The BLUE PLANET Berlin Water Dialogues on 08 November 2023 focused on Closing the Loop: Circular Water Economy. For the fourth time now, the BLUE PLANET Berlin Water Dialogues took place entirely virtual, allowing participants from the water sector from all over the world to join. Watch the Break-Out Session on „Water Energy Management – Recovery of thermal energy/Treatment plants as energy“ moderated by Dr. Bettina Rechenberg, Head of Division III “Sustainable Products and Production, Circular Economy”, German Environment Agency (UBA) here.
Find out more about the conference format here: https://blueplanetberlin.de/
So this breakout session will deal with the topic um how um can we improve Energy Efficiency uh at wastewater treatment plans how can we secure the plant’s own power supply in the long term and additionally um how can we integrate the plans uh in the energy system and I’m looking forward to the
First presentation this will be given by tosa S and she will present a project of the schwat group for an energy intelligent wastewater treatment plant tosar sesh the floor is yours uh hello everybody uh I hope you can hear me now and you can uh see my shared screen uh welcome to my
Presentation about energy intelligent intelligent Municipal wastewater treatment plant it’s a fullscale demonstration plant in a project in Bavaria as um Miss rberg already said it’s the aim of this um project was to uh demonstrate that we are um clo we can close the loop between two sectors uh
It’s water and energy sector uh the first uh the first thing I want to talk about is what is an energy intelligent wastewater treatment plant um um in academics is uh it’s already um discussed a few years and some or most of the parts are already demonstrated on
Uh bigger scale uh wastewater treatment plants but what we wanted to show is that also smaller Wastewater treatments plants can um manage and hand handle energy intelligent Waste Water treatment so what are the uh key factors of this uh concept we have uh uh focused on the
Production side of the energy and we were thinking about what uh energy do we need uh or what are the PO potential energy sources we have so we found that we have we can supply energy from Hydro electric power from photoes from uh CHP unit and uh we can
Um enlarg the storage capacities so that we can um combine all those uh producers and also the consumers by a smart crit uh so we um enlarge the waste water treatment plant by a bigger biogas tank uh storage battery we usually don’t have on wastewater treatment plant and also a
A bigger ter termic storage all those components together have to be managed by a smart grid and supported by a uh software that is integrating all parts of this um the opportunity was or we wanted to take is also to integrate the energy sector and the uh water sector that
Means we wanted to supply our Surplus energy to the local um grid and we call it we wanted to or we want to um operate the wastewater treatment plant in a so-called um Grit friendly mode um next slide I hope that it’s not unmuted can you still hear
Me yes can hear you and perfect then yes um so the wastewater treatment plant the demonstration site is located in schaten book um the design of this plant is um 30 5,000 uh person equivalent that means is uh size um uh wastewater treatment plant
Size 4 plus it’s um not a very big uh wastewater treatment plant but still um uh size four that means a second uh biggest size class of wastewater treatment plants in Germany um uh the design parameters are that uh the future future limits of these waste water wastewater treatment plants are uh
Increased that means uh we have to um we have to uh discharge waste water up to 5 milligram per liter ammonia and 1 milligram per liter phosphorus and also we increased slightly the capacity of um the plant from 168 L per second uh total uh inflow to 220 L per
Second important was to think about energy already from the start also in the planning process so what we or the first planning decision was um to create a wastewater treatment plant with a step feed by a l uh nitrogen removal uh this is also this can be seen
On the plan on the left hand side uh in the uh uh in in the south of the waste water treatment plant and also what we planed from the start is a processed water treatment and in this case a deammonification which means a very low energy
Consumption so the the I the idea the idea was to focus on energy on this wastewater treatment plant and to create the First Energy intelligent wastewater treatment plant in Germany without co-fermentation and in size class 4 the project energy intelligent V treatment plan schaten was funded by the
German um uh Federal Ministry for the environment and many more with um an amount of almost 4 million uh Euro the total project um uh uh costs were are are are about uh 25 million uh Euro um as I already said we created a smart crit um to combine all um
Electrical and um heat uh uh producers with uh the consumers the challenges uh was the spatial te temporal harmonization of decentralized energy production and energy consumption and we hope to uh that the project contributes to a cross- sectoral Energy Efficiency that the plant CO2 emissions
Are to be zero in the end uh as due to the fact that we only con consume R renewable energy and uh we want to totally avoid uh the expansion of national power grid and local distribution networks by serving as a um a net friendly wastewater treatment
Plant in the planning process we started with um simulation of a summer day and a winter day in order to show that all those um um storages we we want to install are uh really necessary and that the sizes of the storages are also um sufficient on the left hand side you see
The simulation of a summer day uh without storage capacity and on the right hand side with storage cap capacity as you can see we result in a very low um um use of the grid total um or the the calculated uh storages we needed are for the intelligent uh wastewater treatment
Plant are listed on the right side of this um table and the conventional waste wastewater treatment would have less a storage as I mentioned before uh what is essential for the management of all those energy um uh stakeholders is a energy management software um the software adapted in this wastewater treatment plant was
Um uh adop was adapted in order to um to have uh completely overview of over all those components um the the main factors are the collection and visualization of all energy data uh we have a realtime observation and an analysis um we have a self adopted continuous forecast of energy demand and
Energy energy consumption or and production on uh three different time scales it’s from a few hours to five days we have a realtime optimization and control of energy management system and we have have a foresight management including um weather weather forecast and uh rain forecast um by using flexible producers storage
And uh even uh new flexible consumers in the next slide you can see um a picture of the con Shon in September 2023 so the um wastewater treatment plant is already in operation also the sludge uh treatment at the moment we are about to uh test the operation of the energy
Management uh system it started in October um in December we will start with the operation of the hydro power plant so that we can say at the moment uh currently we we are operating in optimization mode uh next week uh next year uh we plan the start of the
Monitoring program to verify the project go goes over a period of um one year and in June the official dedication of the W Waste Water treatment plant is planned for the results this is um a bit earlier to confirm all the project goals but what we can uh already uh
Report is that we reduced the energy consumption of the wastewater treatment plan from about 40 kilowatt ston per person equivalent equivalent and year uh in the old um and the former wastewater treatment plant to about 23 kilowatt St per kilowatt hours per person equivalent and year of the new
Wastewater treatment plant and I’m sure uh during the opt oper optimization uh mode we can further reduce this uh uh figure maybe La at last some ideas what are the main challenges and uh of this project and what are the challenges for further applications uh um what we think is one major
Challenge that we up today there are no legal requirements for energy uh consumption um we are looking forward to the new uh European uh water direct waste water directive maybe um this uh uh those requirements will improve um uh what energy management on waste water treatment plants
Already the second uh maybe for some plants the most important uh factor is costs uh we have additional cost for energy in for the energy intelligent infrastructure first of all all um we need highly energy efficient unit units that us usually cost more than uh uh conventional uh energy or or unit um
Whatever pumps or whatever you want then um we have additional units for energy production we have cost for the storage units of course we have the cost for the energy management system uh that might be reduced in the future when it becomes a standard product and we have costs of training
And human resources of course another another challenge is the collaboration between sectors water and energy in this case um on a local basis there is all um there might always be some um uh negative reactions towards the idea because uh then also the energy consumption is reduced and the main idea
Of the uh energy or the the Emu is normally to sell their energy and not to um redu or to sell less energy so one example is also the feed in tariffs that have to be um changed in order to make it more attractive to uh wastewater treatment plants to become um grit
Friendly and last but not not least um this uh is a wastewater treatment plant that is um constructed completely new but the real challenge will be uh to implement those things on Rec reconstruction sites thank you for your attention and the floor is open to your questions yeah thank you very much for
This really inspiring presentation and uh we have uh a couple of questions we have uh a couple of question dealing with the energy sources um so um one question is why uh do you exclude wind power second question is um how do uh you link uh your sewage treatment plant to to the
Hydro power plant and second question to this topic is um how you you link an asynchrone energy source with a synchrome um yeah with a synchrome management of a seage power plant okay I start with the first question um we didn’t ex exclude wind power um it
Was just uh that uh there’s not much wind on this side where we uh where we built the um uh wastewater treatment plant as you could see on the picture we have uh we are very close to the um next no you I I think it’s not shown but uh we are quite
Close to the um to the next uh houses or to the next um buildings uh of of our um uh wastewater treatment plant producers so this is one thing and in Bavaria as you might know we need a a minimum distance to houses of 10 times height of
The wind uh uh Power Plant so this was is one thing the other thing is that we are in the middle of this the in the middle of uh of the woods so uh there’s no not much wind in in in in uh on the bottom
Um Hydro power yeah and a second aspect is that we have enough space for photovoltaic so we didn’t expect to need uh another um uh energy source as we already head and for the Hydro power uh we expect to have about um 6 to 10% of the total energy consumption of the
Wastewater treatment plant from the hydrop power so it covers already 10% of the total energy demand which is quite quite a lot for just a single Source um uh what was the question about the hydrop power again um how it is linked to the um wastewater treatment plant ah
Um it is installed uh right before the uh effluent uh pipe so it’s quite close to the wastewater treatment plant it’s not far away it’s maybe it’s about the distance is about 10 m uh height and uh maybe around 30 m distance so it’s quite closed and it’s
Completely integrated into the uh uh uh Pro or the all all functions of the of the waste for treatment plant and the second was about the synchronization of the power sources exactly uh um it doesn’t it it it doesn’t work independently from the grid that means we
Are we are not um autonomous or completely autonomous we can uh work completely autonomous uh uh by the CHP units so they manage the uh synchronization however um we still do uh uh take uh energy from the git and also we uh we give uh energy to the git the idea is
Not to be autonomous completely uh just we we want to make sure that we do not uh um uh take energy from the grid when the sun is shining and all uh uh and all photovoltaic uh plants are uh giving uh energy to the grid we want to make sure
That we uh take um the energy from from the grid when it when it has energy that’s why we are um um we are operating the CHP units during the night and we are taking the um um the photovoltaic energy during the daytime that’s the basic uh idea of this
System okay thank you very much uh a short last question and then we have to uh proceed um becoming uh climate neutral how you are dealing with the carbon dioxide emission from biogas or the methan emissions from biogas um methane is completely us used in this CHP unit so it’s completely um
Um destroyed and um we do only um we have only C2 uh CO2 emissions at the side apart from uh primary emissions from um nitrogen oxy uh uh nitrogen n Nitro nitrogen dioxide so that’s but that’s all okay thank you very much thank you very much again for the presentation and
Congratulation to this really inspiring project and now I want to handle over to Philip deer and Philip deer presents a digital model to optimize the energy demand of a sewage uh treatment plan Philip DEA the floor is yours so yeah as I was as I was saying
Um so that this presentation here is about how can we incorporate process simulation tools and well the fact that we can actually simulate the process itself um into the operational aspects of a waste World treatment site and how we can use that to optimize um well this
In this aspect the the Energy Efficiency of the site um where I would like to start off is um what zans essentially understands um under the term digital twin or the BW digital twin um where it starts where it ends which tools we can
Use uh and then how we can apply that to the um Waste Water treatment site um so if we look at the entire plant life cycle um you obviously have your engineering at the beginning um where that is plant design feed uh construction um you have the engineering
And the commissioning so how do we uh design the process um what do we build how do we build it which Technologies are we using um and then later on the the operational aspects um how do we operate it what is the the best operating Point um how do we maintain it
Efficiently and uh for all of that we use um well we can offer un tools um what I would like to focus on in this presentation is essentially the the process aspect of the digital twin or the process digital twin um which comes into play um in the optimization and the
The operational aspect and here I would like to um well not show but um highlight gproms as a as a process environment tool process um environment uh simil it as a as a as an simulation tool for the automation system and um the actual automation system which would
Be thematic pcs7 ortic PCS n um and the gproms digital applications where which allows us essentially to couple the offline model with the DCS on site um to enable us to have a real time optimization um now what what are the challenges um that we fa during um well
Wastewater treatment um so generally speaking it is quite complicated uh to run a waste waterall treatment side um because on one hand you have legal requirements for the eent uh which constrains you quite heavily um you have a variant influence um whether that is with respect to load or with respect to
Amount um and the biological and the chemical um Network NW works and reaction networks that that do occur during each of the process steps um is quite complex on one hand and it is incorporates very slow Dynamics so you’ll not immediately see if you’re taking a control action if the process
Is heading into the right direction um what can we do well there are two options um or two solutions really um the first one being an operator training system or OTS um which essentially links the the DCS with the the process model and that enables um well as the the name
Suggests to train operators um on well normal operation as well as during extreme conditions when there’s a heavy rain event or there’s I don’t know something wrong with the plant or aign trips or sensors are failing um whatever it may be um and that sort of allows you
To prepare your current operators but also make that transition from um over staff to near staff um well less painful and and more straightforward um the other option is that you can have the model uh run offline and simulated offline um which essentially comes into play when you
Have a very small size size plant um where we talk about couple thousand couple 10 thousand um of PE um or you can take the entire thing online um have it uh directly connected to the DCS whether as in closed loop uh or within within human in the loop and that
Enables you to run effectively sites that are half a million PE or bigger um in a in a fully automated way um and yeah that should give you enough sort of a financial benefit um to invest into into these Technologies um and yeah as the upside that it also helps um with
Climate um what are the tools that I just mentioned um so we have G promps or goms the simulation tool um that we build our models in and um gproms is the underlying platform that is used which houses all the different um solvers all the different functionalities where is steady state
Dynamic optimization simulation parameter estimation model validation Etc on top of that sits the environment that we use to build our models in um where that is Du costom equation or through pre-existing Library models um and in there we use the the Wastewater Library um which yeah how of all the the
Units that that you would need whether it’s an RA tank um a settler digestor Etc and um given the background from G promps where it’s heavily focused on the chemical industry that allows us to also incorporate uh things like CHP boilers engines um treatment sites for upgrading
The the biomethane or the biog gas to biomethane net grid integration of into the gas Network um Advanced CO2 capture Etc um so all that can be incorporated if if need be um and whatever the site is is sort of allowing and using in terms of Technology um GS itself is the
Desktop tool and once you have the model um in there and you’re you’re happy with the accuracy you can then embed it online into the the digital applications platform or gdap short and um gdap essentially is the interface between the offline model and the DCS where you
Relay tags um relay real time uh data into the model and whether it runs fully automatic um with certain time triggers where it runs an optimization every hour or every day or every week um and directly writes into the DCS where it sets new set points for the oxygen
Controller or for the return slud um or the mlr um or it gets sort of passed onto the operator where the operator then needs to accept the the optimized set point um the the digital twin itself of Waste Water treatment site as I mentioned earlier has essentially two use cases um
How are they build up um on the left hand side you can see the operator training system so here we use the gproms Master model um which is validated and built offline um if it’s then connected to simmit um simmit itself is essentially a um virtual representation or a simulated version of
Your control Hardware um and that allows you to have all the actuators um acting as as they should do in the real world uh but you have it as a digital representation um and that sort of gets in from the OS of the DCS um where there is no direct feedback
Um meaning that you can train operators directly on the system where the control Hardware is simulated in simit the process Hardware or process Hardware but the process itself is simulated in in gproms and it is fed by actual live data for what this scenarios depending on
Whether you want to run run those and um there is no feedback loop essentially so whatever the the operator is doing during during the training on the system itself is not getting relayed back into the actual site meaning that you can train on on a
On an offline ver of your of your actual famp um if we look at the the realtime optimization aspect um we’re using the exact same model as we would during the um for the Opera training system um that is as I mentioned wrapped into to gab or digital applications platform and is
Then connected by an opcua Port um and the opca server hosted in the DCS um and gets data relays it into the into gproms it runs the simulation runs the optimization and um sends back optimized set points for for your different controllers uh depending on whatever aspect you want to optimize um
Yeah the the most normal case is energy optimization where it runs um the optimization for all the different pumps for all the different um all the different blowers and it will also keep in mind the um the E quality in the permit L as a constraint so
You’re you’re sure to not run over your your permit L and well get H um something to note here is that uh the real time optim ization or the RTO is completely V diagnostic um so it doesn’t have to be a cement um BCS could be whatever Rockwell or or a Scara
System y uh as long as there’s a direct connection via opcua possible into into the the control Hardware um that’s perfectly fine um how does the digital twin fit next to your your real uh plant um so on the process you have your actual site um in
The digital process T you have the model and G prompt if you go to the field level you have your actuators and sensors of the what are represented by the physical Hardware uh in your in your plant um and they are used or simulated by a simmit um the automation level is
Exactly the same for both um where on a zement uh on Z Hardware you have your sematic um and on a on a virtual level you have well the simit virtual controller which has a onet to one integration uh into into Z product and then you have the HMI level where the
Operator sits and an additional twin on top of the HMI um where there’s a direct feedback link uh or feedback loop you have the the digital application uh platform which again how the Master model receives up to-date and real time uh data that some data validation already built in uh cleansing quy sensor
Data Etc and um gives you a whole representation of the current state of the plant and with that you can then look at soft sensing how do we take plant measurements and infer calculated variables um based in real time so you’re not lacking information or you get additional
Information an operator and you don’t have to wait for the next Lab measurement the next measuring campaign Etc uh and that helps you to sort of monitor the process as well as well diagnose any any problems um on the other hand you have the realtime optimization where you’re even able to
Read in the electricity prices if you have a CHP plan and you look at what might actually be worth now do I should should I store my bio gas for a later point because electricity is pretty cheap at the moment or should I use um
The bio gas now in a in a CHP and um save the money I would spend on importing electricity um and if you have all that information you can then um run the optimization and get recommended optimal set points for your control variable um how does that look
Like in the flash or in the ones and zeros um so so what you have here is the gproms environment um all the different libraries available to you very easy flow sheet and you just drag and drop the different um units connect them up specify them um ideally that’s it
Um similarly as as simple as it is you define your objective function you set the decision variable um so which variables do you want to control which variables do you have a handle on um your constraints eff quality and that that’s pretty much it from the the gon side of the process modeling
Aspect of it obviously there’s a lot more detail involved uh and you need to validate it you need to run measuring campaigns you need to sit through data make sure that the model is as accurate as it as it can be or as it should be as
It needs to be um for the operator training system um yeah simit takes on the control Hardware so in here you would have your um your different controllers you would have uh different actuators and sensors and you would link that directly to G promp so you have direct mapping on
On an opca level where the calculated variables from G promp are relayed into simit and the um control actions implemented in zimit get relayed into into G promp and all of that is then displayed in the in the OS in the operator station of the DCs so this is
In PCS Neo um for the operator training aspect you have this handy feature where you can run the simulation in well up to 3,000% real time um which allows you to deal with those slow Dynamics um so that’s why those numbers are are flickering quite quite VI quite
Vividly um and that’s a representation of the gums model that I just showed a minute ago um where you have your d d nutrification tank the nutrification tank some dosing your air blowers your mlr and your return flood um your soft sensors your secondary settler tank and
You can change all the values all the set points manually through a face plate um so if I would increase the the influence volume here um you would see that all the controllers TR to compensate for that um that change in in Inlet and you have the current set
Points um of our three controllers so we’re trying to control the um obviously the the air Inlet um we are controlling the mlr and we are controlling the return SL and those are our three set points now if you wanted to start um gab or the the real time optimization we can just
Do that very easily um that will just pop up another command window um and you can see that it’s running and the calculated variables will then be put back or sent back um it will take a minute um there and once there’s a sent back there we go
It’s finished now so we have new set points yeah let finished now um so we have the optimized set points that we just calculated given the model given the constraints that we that we set um and you can see that it’s it’s quite a hefty energy saving potential that we
Could realize um so what I want to do now is take those set points and incorporate them into the model and if I go back you can see that new set points have been um adopted by the PS and obviously it will not be as quick in
In real time or on site actually um but the the PS will control onto the new uh set point and one once the actual process the real world process has status uh you can enable manual input again and now your your operator is again free to make
To make any changes here um and deal with the next set of problems or operating points um until something changes that is worth making a new optimization running a new optimization the backround uh and then it just Rin and repeat from there on um and yeah so in this example we saved about
20% um which yeah is quite good um and if I go back to my presentation we have one more slide so where was this applied um obviously we have quite a few um projects where where we’re working on um unfortunately we’re not able to share most of them uh and
The results that that were achieved um so here’s an example that we we can actually um show and so this was a um wastewater treatment site in the swi in Switzerland U had about 20,000 PE and the question was how can we run the or operate the the site more more
Efficiently um the problem here was that the influence was a mix of indust indust SL industrial Waters and communal waste waterers um which meant that at various times throughout the day or throughout the week um the inflow of or the ratio between communal and Industrial Waste Water was was quite off or quite
Different meaning that there was several different um operating points that that needed to um be optimized and during the base case for example um the the model was deployed offline um and was used as a decision support tool by the operator um and the operating potential was was
Quite quite large um again around 20% annual electricity consumption without having any negative impact on the water after in quality um mainly because well some of the the nutrify cells could could be switched off during during base operating Point um and and even at the high level
Windows and there was quite a lot of uh Industrial Waste Water coming in that is very highly loaded in in ammonia um you could still reduce um the the operating facility the the amount of sales that that you needed to get to the E quality that we were allowed to to
Discharge um so yeah 20 20% quite on the high side um and yeah other than that um I think it’s just an example that you can you can use it on on small sites you can use it on large sites uh you can use it offline as a decision tool or you can
You can run it fully online in open or close loop um so that there should be a use case for all different sites and and yeah all different sizes and different applications of waste water sites and that is me I hope I didn’t overrun okay thank you very much Phillip
For the demonstration of this really impressive tool um there’s only time for one question um and the question um we we’ve heard in the first presentation uh that an important topic is the integration of uh the operation of su treatment plans in in into the
Grid that means to to come to a grid friendly uh operation uh could the digital twin um also support uh yeah an INT intelligent operation um of the sewage treatment plan so adopted to the GD relief or or feed in when required um in principle yes um usually the limitation stems from
Not having enough data um and therefore not being able to accurately represent whatever you you want to simulate and therefore represent um however there there is the option that if you have control on your on your digestors or your your thp or whatever it it might be
That reduces energy on site um and you can you can get a grip on your um storage capabilities then that should allow you to have uh at least on a very basic level an understanding on when would be a good point in time to either feed in
Or or import or export energy um and then that needs to be coupled with a with a wider grid level simulation I guess um but that that was not something that we’ve done already uh or that we are currently even planning on doing um nevertheless I think the the base blocks
That you would need to enable that are there um so you could theoretically do that yeah okay thank you very much thank you for the presentation thank you again tosa and uh we are now at the end of of our breakout session and I hope um I can
See you soon all on the main stage in five minutes so thank you very much for the presentation and for the contribution and fruitful discussion see you soon on Main stage