The Next Big Opportunities in Energy Storage

Now, I’d like to introduce today’s featured presenter William Chueh. Will Chueh is a assistant professor in the Department of Material Science and Engineering and a center fellow at the Precourt Institute for Energy at Stanford University. Professor Chueh has received numerous honors including the MRS Outstanding Young Investigator Award in 2018, The Volkswagen BASF Science Award, Electrochemistry in 2016, Camille Dreyfus Teacher-Scholar Award in 2016, and many others. In 2012, he was also named as one of the top 35 innovators under the age of 35 by MIT’s Technology Review. He received his BS in Applied Physics, MS and PhD in Material Science from Caltech prior to joining Stanford in 2012. Now, I’d like to turn the floor over to Will. Thank you Joe for the introduction, and, uh, it is my great pleasure to, ah, give you a brief presentation on the future of energy storage. So, if you look at the cost of lithium ion batteries, for example, since 2010 have declined by almost 80 percent, and the trend is only increasing. On the other hand, if you look at the cost of generating electricity renewably, that cost has even gone- gone down even more. For example, for wind electricity since the ’80s has gone down by 90 percent and for solar that has gone up by 99 percent, um, in the past two decades. And if you combine the decline costs of storage and generation of renewable electricity, that leads to two very interesting questions. The first question is in the sector of transportation: So, what are the remaining research and development needed to fully electrify and convert our internal combustion engine fleets to electric vehicle? And a greater question and a bigger challenge is for the grid: So, can we combine this low cost renewable electricity and low cost storage and compete on cost with conventional electricity generated from fossil fuel purely on cost? So, these are the two question I’m gonna try to, um, give additional details. I will discuss them both from a business perspective and also a technology perspective. And let me just state the take home message, which is you have to look at the problem both from the business side and the technology side. So, I’ll go over some details in the next 45 minutes. So, let me start with a plot showing you the tremendous growth in EVs. Uh, this is plotting the millions of vehicles, um, for the past, um, eight years from 2011 and projection into 2019, and you can see we sold our first million electric vehicle in 2015 then the second million 2016, and now we have sold over 4 million passenger EVs, and within the next year we’ll add another million to that. The numbers are big but they are very small compared to the global industry of one billion vehicles. So, that tells you some of the opportunities going forward in terms of fully electrifying, and I don’t think I’ll have to highlight the importance of battery technology for transportation. So, I think most people can appreciate the growth of EVs. Perhaps a harder problem to see is how energy storage is impacting the electrical grid. And here I’m showing you the global installed capacity for renewable generation. So, this is showing in green for wind and in yellow for solar. And you can see over the past 18 years the increase in this all capacity of wind and solar has just being enormous. Now we’re approaching approximately one terawatt of installed capacity for wind and solar. So, the electricity grid is changing very quickly and one of the key challenge is how does the grid cope with the generation of what we call variable electricity? So, these are not resources which is available all the time. The sun doesn’t shine all the time and the wind doesn’t blow all the time. So, how do we think about changing the grid to accommodate for this variable generation of electricity, is one of the key challenges going forward. So, let me talk about energy storage, um, in the context of the electricity grid. Ah, these are plots showing, um, the utility scale energy storage installation that is forecasting from, ah, 2017 all the way to 2026. And here on the left, um, we’re looking at, um, the energy, um, storage in terms of utility and on the right, it’s the installed, ah, capacity for distributed energy storage. So, utility means very concentrated, so these are power plants style energy storage technology, and on the right, these are distributed or a home based or, um, a business space energy storage. And I won’t get into the details of this plot which is being distributed over, um, many countries, but the key aspect is the trend. Energy storage is going to grow very rapidly both at the utility level and at the distributed level. And the reason is actually quite straightforward. One easy way to deal with variable generation of electricity is to be able to buffer it with energy storage technology such as batteries. So, one of the key driver for energy storage is policy and, um, being here at Stanford- let me start with California. So, there are several very notable policies that is driving the investment in this area. For example, the mandate to have 1.8 gigawatts of installed, ah, storage by 2024, and more recently to go completely carbon free electricity by 2045, that was just announced two months ago. You can see similar initiatives in the policy area in Arizona, New York, Massachusetts, Texas, and many places in the US and globally as well. In the US, um, the state level policy, it is currently what is driving energy storage on the grid level. And this is going to be a very significant factor going forward in the growth of the industry. So, lemme give some specifics on battery technology. Battery, it is not the only technology that can be used for grid storage. I will talk about a variety of them, ah, later in the webinar, but just show you one example. Um, this is for battery based technology and you’re basically seeing almost zero installation in capacity prior to 2010. But because the cost of lithium on battery has gone down so much, you can see that the declining cost is almost matched with the installed capacity which has really taken off since 2010 as well. This is a great start but we’re still very far from matching this to the one terawatt generation capacity I showed earlier from solar and wind. So, this highlights the challenge we have to deal with, and I will talk about a little bit some of the technological and the business drivers behind getting the storage capacity to match up with that of generation. So, let me talk a little bit about the application of energy storage in the context of the grid, so we can all appreciate how energy storage can help with transportation in terms of electric vehicle. The grid is a much more complex system. And you can think about the utility of storage in the contexts of the duration of storage. So, how long do you have to store electricity for in order to tackle some of these various applications? And what recent applications? Each application has a revenue associated with it. So, this is driving some of the technical economic behind storage on, for example, whether utility companies will incorporate storage or not, or do they go with traditional means of providing electricity, for example, through natural gas power plants. So, let’s start at the bottom. On the level of seconds, you have things like startup of the grid, so you have to provide some initial energy if the grid was down and you have to restart it. The second one is frequency regulations. If you look at your electricity output, it is very stable, okay, but, ah, that only happens when demand and supply of electricity are perfectly matched. And sometime it is difficult to match it so you have to have some amount of buffering so that your electricity will come out at the right voltage in the right frequency, so 120 volts and 60 hertz in United States. Then you go on to minutes. So, these are what’s called reserves. So, to accommodate for variations in the grid, you have to have electricity that you can provide, um, on a reserve basis on the order of minutes. Then keep moving up, you have what’s called peak shaving, and this is corresponding to the fact that there are demands of electricity that is high and low in the day, and when the demand gets too high, ah, this puts a lot of stress on the system, on the grid. So, being able to shave that peak, ah, to decrease the peak requirement of electricity from conventional sources is important. And then if we move up even more, in the purple box we talk about curtailment. So, this is where there is so much renewable being generated, for example, from solar-wind, you can’t use it all. And if you can’t use it all, you have to start, ah, either dumping the electricity away or you have to think about basically curtailing it. And this is a- a significant revenue loss for the electrical, uh, electricity grid companies as well. And on the very very top right, you have seasonal or long duration resource shifting. I’ll talk a lot- a little bit about this as well. So, you have different outputs of electricity from solar, uh, throughout the year. So, we are- have high solar outputs in the summer and lower solar output in the winter, but the demand is actually, ah, somewhat uniform. So, being able to shift- the ability to shift the electricity between seasons could also be very important and another important source of revenue. And if you look at this plot from left to right, it also shows you how important it is dependent on the penetration of renewable. So, when that renewable penetration- if you want it to be a 100 percent, you have to talk about seasonal storage, but if you only want it to be 10 percent, then maybe we can get away with just peak shaving. So, let me show you one example, ah, of a typical, ah, one week electricity, ah, generation here. So, here we have green, it’s the generation of wind, and we have here, ah, in yellow at the top line that is the total electricity load. So, if you look at the green here from left to right, you can see it’s highly variable. It can be almost zero at times, and it can be very large at other times. And if you look at the low profile, ah, the up and down, ah, reflects a cyclic, um, demand of the day, ah, higher during the daytime and lower during nighttime. And where you have a peak demand and low when output is where you have to provide electricity from something other than wind, and this is where we turn up our natural gas power plant. And on the other hand, if you have a high supply, a low demand, this is where we have to curtail. Which means that, ah, the grid- electricity grid companies are paying, ah, wind farms to basically turn off the wind turbines because they cannot take the electricity and that sometimes can result in negative pricing, so the grid companies are paying, ah, for electricity not to be generated. And if you look in the middle of the plot highlighting to potential, um, um, areas, you can, for example, shift the resources. So, right around say here February 24th at, ah, noon you have high wind generation and then at midnight you have low wind generation. So, if you can move the electricity around by that, ah,12-hour, ah, sorry that 20, uh, 36-hour period, that can also provide additional revenue to the grid companies. One other application is frequency regulation, ah, this slide is showing quite a bit, but let me just tell you that regulation in frequency happens on a very fast timescale, so these are typically on the order of seconds. And so these various plot basically shows how effective it is to match the frequency of the electricity to the, um, between the demand and the supply. On the top right, you have a case where you have battery based buffering, so the two curves are matched perfectly. So, there is no lack behind between the supply and demand. But if you go to say something like a hydro storage, so this is using a water by pumping up the hill and down. You have a slightly greater mismatch, um, because of the lap time in providing the electricity. And then if you look at a- a turbine based combined cycle on the very lower plot, there you can see that, ah, the lack is even greater. So, this tells you that the ability to maintain a very stable grid has to do with how fast you’re able to buffer the electricity and battery provides a really good way to do so, and this can directly generate revenue, ah, for the electricity grid. So, here’s one example. Um, many of you might have heard about, um, the installation of a 100 megawatt, uh, 129 megawatt hour battery system in Australia that was installed by Tesla in, ah, 2017. The main application of this battery system which is based on lithium ion batteries is to provide for frequency generation, which means it’s used to buffer the electricity, ah, so that the frequency is held constant. So, 100 megawatt or 121 megawatt hour is not a very big number, but it is very exciting to see even a small installation like this can have a substantial impact on the, ah, type of storage used for frequency regulation. So, if you look at the right, the plot of the first, ah, bar shows you Q4 in 2017, right before the installation of this battery system. You can see battery was a very small percentage of what’s used for frequency regulation of the grid and is mainly powered by coal and hydro. Okay. So, you’re using coal and water to provide the buffering needed to regulate the frequency of electricity in Australia, but just the next quarter, so three months later, you can see battery grew from about one percent to over 10 percent as the means to regulate the frequency of electricity. So, a small installation like this one can have a huge impact, and this is for Australia as a country. So, this is very exciting to see how a small battery system can have such a large impact for frequency regulation. Another application that I alluded to earlier is peak shaving, so let me get into some detail here. So, peak shaving basically reflects the mismatch between the consumption of electricity and the generation of electricity. And here, because the generation of renewable electricity is variable, you have to buffer the differences between the two. So, one application of peak shifting is as follows. You have the top of the demand at about 06:00 PM, this is when everyone goes home, turn on electricity maybe turn on the AC, uh, if you’re in the summer in the southern states in the US, and a nighttime, for example, the electricity demand is much lower. So, one idea is that you can generate, ah, excess electricity when you have lower demand and shift it by about 12 hours to higher demand. One particular application could be in storing wind electricity, so wind often blows at night. So, if you’re able to store that, basically you can charge your battery at night and then you discharge battery during the day when the demand is higher, this will lessen the demand on the electricity in the peak hours. So, the ability to shave this peak can stabilize the grid and make it more economical for the grid to be operated. So, one example of this peak shaving is the California duck curve. Some of you may have heard this in the popular press, that duck curve basically shows you the shape of the generation of nonrenewable electricity and how it varies throughout the day. So, if you look at the plot on the left, it is showing you the generation of electricity from non renewable sources. Okay. So, these are primarily your natural gas power plants in California for a typical spring day. And the curve on the top in gray is the actual in 2012, and as you go down it shows you the prediction to 2020, and the prediction was made in 2013. So, you can see that because the solar insulation was still quite low back in 2012, so your generation of nonrenewable electricity is fairly flat during the day. But as time went on, the solar installation increased in California, which means you’re outputting much more renewable generation during the daytime. And as that renewable generation went up, that means that generation of conventional electricity from natural gas and others have to go down. Okay. And that’s why from 2012 to 2015, you’re seeing the formation of the site profile of a duck. Which tells you that you are needing less electricity during the day, because you have more solar electricity to work with. What problem does this give you? Well, you have to shut down the generation of nonrenewable electricity during the day, and you have to turn it backup right around 4:00 to 5:00 PM when the sun is beginning to set and that gives you the neck of the curve. And if you look at the number here, we’re talking about the ramping of approximately 10,000 megawatt over a three-hour period. This is a tremendous challenge and adds a stress on the grid, and this was the prediction back in 2013. How good were the predictions? Not very good. In fact, what we are today, we’re already way past the 2020 projection in terms of the duck curve. So, we’re very close to the bottom of this curve, which means we are shutting down a tremendous amount of conventional electricity generation during the day time and we have to turn it back up over a three-hour period. And the plot on the right just shows you how much ramp we’re talking about, 10,000 megawatt. To put some context, 10,000 megawatt is half of the Australia grid. So, in California we have to turn up electricity that is equal to half of the grid utilization in Australia. So, this duck curve it’s a tremendous consideration when it comes to technologies that can mitigate the duck curve. Let me show another example of curtailment. So, this is a projection in 2020 for the western grid in, in United States showing how various form of electricity- uh, va- various forms of generation, uh, for electricity. So, at the very bottom, you have nuclear. So, that is very steady over time, so it doesn’t vary. So, the power plants are running uniformly. Then, you have coal and hydro. So, you can see coal actually goes down during the day, and the reason is very simple. The grid company turns off coal first because it has the largest carbon footprint, and the reason you turn off coal is because during the daytime, you have the increase in solar, and that is captured in yellow there. So, this essentially shows you the different types of electricity that is being generated day to day. [NOISE] And if you look at the very, very top, there, you can see curtailment. So, there, you have so much electricity that you couldn’t turn off enough of the coal and other conventional generation, so that you have to curtail the renewables by paying the, uh, uh, renewable generation company not to generate, to turn off their resources. And this is a hit on the economics because the grid operator are paying money to the renewable generators and the renewable generators are not able to generate the revenue they originally wanted to generate when they installed the equipment. So, this shows you how dynamic the grid is. It’s really incredible to see the ups and downs of various generation of electricity throughout just one day. [NOISE] And as another example of resource shifting, these plots are showing you the resource and demand profile on the left for California and on the right for the United States for the entire year, okay? So, you can see the demand is fairly flat, okay? In California, it’s fairly flat. In the US, it’s a little bit, uh, less flat, and that is the or- um, the orange curve there. But what is very interesting is, if you look at the renewable electricity, so that’s in blue on the left and on the plot on the right is separated into both wind and solar in purple and yellow respectively, and there, you can see a significant mismatch, you can see that we have more generation during the summer months and less generation over the winter months. So, the opportunity here is, can we align our resources with our demand by shifting electricity around seasons? So, can we take some of those electricity generated in the month of May to August and we can shift that to the consumption, um, in the winter, okay? So, this is something that is not yet possible, but seasonal shifting of electricity has a huge opportunity, and as I commented on earlier, we want renewable to penetrate the grid completely, we have to tackle this challenge. [NOISE] These are few more plots of simulations. So, what this is plotting is basically how much capacity overbuilding we need. One means no overbuilding, and two, three, four means we have to overbuild the solar and wind by a certain amount, and the X axis is showing you the amount of renewable penetration. So, zero means no renewable, a 100 means completely renewable. The black curve shows you, um, the amount of, um, um, overbuilding [NOISE] if you have only- uh, have no storage, and as you go to the lighter green curve, it shows you what happens if you have storage one day, four day, and one month. So, as you increase storage, then you get to decrease the amount of overbuilding of solar and wind, and the reason is simple. If you don’t have storage and you want a certain amount of renewable penetration, that means you have to have more than you need. So, then, you can generate enough solar and wind when the sun is not shining or when the wind is not blowing. But if you have storage, then you’re able to shift the resources around so that even when the sun is not shining and the wind is not blowing, you can still use the renewable and that decreases the amount of overbuilding that is required. And it’s very interesting to see [NOISE] the various cases. The top curve shows you what happens if you purely have solar, okay? Because solar is intermittent on a daily basis, this highlights the importance of storage. So, if you want 50 percent penetration of solar in the United States, that means you need to have storage. Without it, you can only get to approximately 50 percent penetration when it becomes impossible to supply, and this is simply reflecting the fact that the sun doesn’t shine at night. But if you look at the bottom cases, if you go to 50-50 mixture of solar and wind, the curve shifts to the right, which means for a given amount of penetration, you need much less overbuilding, and this reflects the fact that wind is intermittent on a different timescale, it’s intermittent on, uh, a day- uh, several day basis. So, if you mix just the right amount of wind and solar, you’re able to have less overbuilding for a given amount of penetration, and if you go to 100 percent wind on the bottom, that also showcases as well. But regardless of what scenario you consider, you need to have storage in order to reach high degree of penetration of renewable electricity, um, in the grid system. [NOISE] So, now, let me talk about the alternatives. So, we talked a lot about the various application and the business need for storage, whether it is seasonal storage, whether there is daily storage, or peak shaving. Energy storage is not the only way to do it. There are a lot of competitions that is already available today, and probably one of the most attractive alternative is natural gas peaker plants. So, a peaker basically reflects the fact that these natural gas power plants are not turned on all the time, it is turned on when the demand of electricity increases, for example, around 3:00 PM when the generation of renewable goes down. And this is basically showing you the cost of electricity per megawatt hour shift is. So, this is talking about, if I have to move one megawatt hour of electricity around for a certain amount of time, what will be the cost? And you can see actually, technology like lithium-ion battery is high on this plot, okay? And using, um, a natural gas peaker plant is actually very cost effective. This is based on today’s cost of lithium-ion battery technology. But look at the same plot on the right. In 2030, the projection is that lithium-ion battery will actually become less expensive than natural power- natural gas power plants. So, this is showing you the importance of continuing to push the cost of battery technology so that it can compete with other ways of shifting electricity, for example, through peaker plants on cost, okay? And I want to emphasize this one more time, is that for something as enormous as the grid, techno economics is the principal driving force, you have to think primarily about cost. [NOISE] So, now, let’s get into the technology aspect. So, if we want to shift energy around, whether it is on the order of an hour, three hours, a day, or across the season, there are many different metrics that must be satisfied at the same time, and this plot here shows you just a few of those metrics. And I’ll walk you through from the top, uh, starting with calendar life. So, calendar life basically tells you, for a particular energy storage technology, how long can you use it for. It doesn’t matter how many times you use it, it’s just, um, a calendar. So, is it 10 years, [NOISE] 20 years, and this determines the cost of electricity that is amortized over the lifetime of the technology. Going around in the clockwise fashion, the next one is cycle life. So, this is how many times you can cycle the battery. So, if you’re storing electricity on a daily basis, then you will need to cycle at about 300 times in a year. Uh, if you’re doing frequency regulation like the Australia example I showed earlier, then you’re doing multiple cycles or tens of cycles a day. The next one is energy cost, so this is basically looking at how much does it take to store a certain amount of energy, so this will be the dollar per kilowatt hour that I have shown earlier. The next one is energy density, so this is how heavy the technology would be, giving an amount of energy stored. The next one is power cost, so this is looking at the cost of delivering electricity at a certain rate. So, for those of you not familiar with energy and power, uh, in the context of an electric vehicle, energy tells you how far you can drive the car so that’s the range, and the power basically tells you how fast you can either charge the car, or how fast you can accelerate the car. So, there are costs associated with both. Uh, keep going clockwise fashion, ramp rate, so this is reflecting how fast you’re able to charge and discharge the particular energy storage technology. So, we talked quite a bit about the duck curve, so that particular application will require you to be able to charge and discharge the energy storage technology over hours. But if you’re talking about frequency regulation, that period can be quite a bit shorter as well. Round-trip efficiency, this is basically telling you how much of the electricity you put into the technology you can actually get out. So, 100% round-trip efficient system that means every bit of energy you put in you can take out, and this also determines the cost of storage as well. Safety is very important. Um, you want to have an energy storage technology that is safe over the lifetime of the system. And this applies for both large scale utility installations and also home or transportation. Temperature range is important, and we have many participants here from different parts of the world. Sometime it gets very hot in the summer, sometime it gets very cold in the winter, to have the technology that perform over the broad range of temperature and those of you who have electric vehicle can appreciate the fact that in the winter months then your performance is decreased, for example, for lithium-ion battery and that highlights one of the weakness of that particular technology. And the final one is the volumetric energy density so, this is how much you can store per size. Okay. So, that in combination with the gravimetric energy density basically tells you how big and how heavy the particular technology has to be. The challenge with energy storage technology is that you have to satisfy many of these at the same time. And the plot in the middle basically shows you two example technologies in which you don’t have complete, um, meeting of all the metrics. Okay. So, there’s no silver bullet, and depending on the application, you can weigh these metrics differently. So, for example, for transportation, gravimetric and volumetric energy densities are very important. But for grid application, the calendar line is very important. So, as I will talk about in the next uh, 10 minutes or so, is really thinking about what use case you will like a particular technology to serve, and depending on the use cases, the technology will be different. And there is no silver bullet, so this is why we need a wide range of energy storage technology to meet the different type of applications that we have. So, this is my personal view and costs of vacation of energy storage technologies, I divided them by the physical storage mechanism. So, going from the left we have electromagnetic. These will be technologies such as supercapacitors. Uh, we have thermal storage, so this is storing energy in the form of heat. Uh, this is most notable, um, in the form of molten salt that is used for concentrated solar power, whereby you, uh, concentrate solar light, and you generate heat, and the heat can be stored by heating something up like a molten salt. The next one, uh, is probably the most, um, the one that has received the most amount of attention is electrochemical, and th- the poster child of electrochemical storage is lithium-ion batteries. But you also have a wide range of other battery technology, lead acid, flow battery, and so forth. Further on the right, we have mechanical, uh, and the poster child for mechanical is pump hydro storage so you take electricity- you take water and you pump it up the hill, that’s how you charge up a pump hydro system and then water goes down to the hill. This is how you get electricity out. You can do this not only with water, but you can use air as a medium, and you can also use other things, ah, as well, um, masses, rocks even by moving up and down the hill. And the final one is chemical. So, we have also the opportunity to take electricity and generate chemicals such as hydrogen, and then we can combust it subsequently, and then to make electricity again. So, these are the various aspects. I don’t have that much time today so I’ll only briefly highlight a few examples, but I will like the audience to appreciate the wide range of technologies that being investigated today at universities, national labs, and industry to try to get this work at the scale I alluded to earlier in my talk. So, let me look at the installation capacity by energy storage. By far, the technology of preference today is pump hydro. Okay. An overwhelming amount of energy today are being stored by pumping water up and down. If you look at other forms of storage; thermal, electrochemical, or other it’s very small, it barely shows up on the plot on the left. And the plot on the right basically shows a blow up of those three things. Okay. So, uh, technology other than pump hydro. So, there you have a significant participation of thermal storage, so storing energy in the form of heat. Lithium-ion battery is the biggest one now in the electrochemical area, and you also have other installation in mechanical for example, storing it in compressed air, or by turning the wheel around and spinning it. The opportunity for the future is, how can we develop different technology to reach the same level of economy as pump hydro and ultimately, to exceed it? And the reason why pump hydro has limitation is because, it’s geography dependent, and water resource is always a consideration as well. Let me get right into lithium-ion battery. This is something that has really captured, um, the attention of the public in the past decade because of the rapid, um, explosion of electric vehicles. A battery technology is actually fairly straightforward. You are shuttling, um, atoms such as lithium between a low-energy reservoir and a high-energy reservoir. So, in a manner of speaking, this is not very different than pump hydro except the energy carrier is not water, and pumping up a difference in height, but is really, ah, a lithium that is being pumped around a difference in chemistry between the two side of the battery, which we call the negative electrode sometimes called the anode, and the positive electrode sometime called the cathode. So, the ability to move, um, this energy carrier lithium around the two electrode, um, is how a battery works. One of the most attractive thing about a battery is how inexpensive it has become, so if you look at the table on the top right, it shows you the various, um, uh, parameters for the technology. Currently we are looking at a grid of two gigawatts of installed capacity. Typical duration now based on the economy is about one to four hours for grids, and, uh, is already serving the need of electric vehicle, the lifetime is about several thousand recharge cycles, and the calendar life could be on the order of about 10 years. Round-trip efficiency is very high, about 90% of the electricity you put in, you can take it out. The cost is extremely attractive today for transportation, but it is quite far from what is needed for the electrical grid. I didn’t get into the economics of the grid, but if you want to have a significant penetration of electricity storage using battery, you need to look at significantly under $100 per kilowatt hour for diurnal storage, meaning across, um, uh, a date. And if you want to look at say storage across several days or a week, you need to reach about $20 a kilowatt hour. So, we’re still about 20X off to reaching that goal. So, new technology is certainly needed. And probably what is most impressive in lithium-ion battery is the scaling exercise that we have went through. So, the plot on the left shows you the dollar per kilowatt hour on a battery pack level, has gone from over a 100 per kilowatt hour just in 2013 to now slightly, uh, anticipated to below a $100 by 2025, and approaching $50 per kilowatt hour. So, this is like truly attractive, as the cost goes down new applications will come up. So, as the cost down then you can store electricity for longer periods of time for example, for the electrical grid and the cost of transportation for EVs will also go down as well. And the costs decline will be matched with the growth in the installed capacity and that is what’s shown on the right. It is truly exciting to think that the lithium-ion battery market should in the next 10 years reach over $1 trillion worldwide. This is a plot showing you the companies which are involved in manufacturing lithium-ion battery. A majority of these companies are in China, some of them are in Korea, and elsewhere in the world as well. And most company have been announcing new capacity, so this plot is already outdated, and we’re seeing numbers even larger today in order for the industry to meet the demands for transportation and increasingly for grid storage application for lithium-ion batteries. So, what is the ultimate limits of the cost of lithium-ion battery? So, I showed a projection down to $70 and $50 a kilowatt hour. So, fundamentally, what sets the floor of the cost of battery technology is the material cost. So, this plo- uh, bar shows you the projection of costs from 2015 to 2025. So, you can see, um, the cost of manufacturing, the cost of assembling the batteries into a pack will continue to go down, but the cost of the material will not decrease dramatically. So, the material costs will set the cost for it. So, basically if you look at the projection going forward to 2025, the material costs will dominate the cost of the battery pack. So, this is showing you the trend in the elemental cost. Um, this is the sum of the three expensive component in batteries are cobalt, nickel, and lithium, and you can see because of the rise in demand, you have now a sudden rise in cobalt price over the past few years, and this is going to be a significant constraint. So, one opportunity is, can we design new battery chemistry that do not use these expensive elements? This will be one pathway toward lowering the costs for which is set by the material cost. Another thing to think about for batteries is cost that is level-wise by the lifetime. So, the cost I’ve been showing you so far, is level-wise only by the energy. So, dollar per kilowatt hour, but different battery technology have different lifetime. So, for transportation, we need maybe a thousand cycles. So, if you think about a uh, car with a 250 mile range, 1,000 cycle means it’s a 2- 20 um, uh, 250,000 mile, and that likely exceeds the lifetime of cars in most uh, uh, use cases. But for grid application, we need much more, we need thousands or 10,000 life, uh cycle life, and when you start dividing the cost by the lifetime, the equation changes. So, for example, this plot shows you different technologies uh, in terms of the cost per kilowatt hour and the cycle life. Some of these technologies are not used in transportation today because the energy density isn’t high enough. But if your primary consideration was not the energy density, say you build a big, big battery system in the desert or elsewhere, then you have to actually worry about costs more than the density, and the plot on the right that shows you once you levelize the cost by the cycle life the battery, then different winners emerge. So, the present technology we use today are actually fairly high in the levelized cost because you are achieving your trading the lifetime of the battery for high-energy density. But for certain use case like the grid, you may want a lower energy density, but a much longer lifetime. There are new chemistries out there that can potentially deliver that combination of metrics. So, this is something else to think about too in terms of developing the white chemistry and the right technology that will hit the needs of the market. One other application that is emerging is second life batteries from electric vehicles. So, as we have more and more electric vehicles on the market, eventually, we’ll have them coming off um, electric vehicle. Maybe they won’t be good enough for transportation, the range is too short, but we can still take them out and use them in grid application for example, by using them to buffer electricity for solar. This simulation basically tells you that by um, the beginning of um, 2020, we will begin to see significant amount of battery come off electric vehicle, and this second life use of batteries represents a very large market as well, and economics is not fully understood and there is a lot of opportunity in thinking about, how do we effectively take batteries out of electric vehicles and repurpose them for other applications such as the grid? So, I’m out of time today. So, maybe I will just say a few more words. Um, I didn’t have time to get into the specific technologies here. The take-home message here is there are a lot of different technology, and each one of them will have a different um metrics. It’ll will not meet all the requirements, there is no silver bullet. RND is ongoing for everything I’m showing you here and more. Let me stop with this um, take home message here, which is looking at the big opportunity. My opinion is that we cannot work on technology without considering the business needs and the use case. So, I spend about half the talk talking about the various use cases, how the uh, grid companies could be using battery to generate revenue, we can be looking at transportation as a way to replace internal combustion engines, we talked about the economics behind it. On the right-hand side, we have the technology. I only talked about electrochemical today, but there are many more. Really, it’s about matching the two, finding the right technology for the specific business cases, and identifying the need to further develop technology to meet certain technical economic goals. I think this is what is needed to really see the penetration of renewable electricity, and the role that energy storage will play. So, let me stop here and thank you very much for listening. Now, we have a few minutes for William to answer a few of your questions. All right. So, the first question is, what are [NOISE] the challenges through recycling batteries at a large scale? What are the associated environmental impacts? How does techno- how is technology reducing the environmental footprint? Uh, this is uh, a very loaded question I must say. So, let me answer the last question first. The environmental footprint is not an easy question to answer. You often, you know, see in, you know, um, on the license placing, this is a zero emission vehicle. Uh, that’s not entirely true. It depends on how the electricity was generated. And it also, you need to think about how much emission went into making the technology. So, here at Stanford and elsewhere, folks are doing analysis on the life-cycle footprint of these technologies. So, as you compare different technology, you have to think about what is the emission that went into the extraction of the raw material? The manufacturing and the assembly of the technology, and that has to be considered. There’s no simple answer and one can simulate what the carbon payback time uh, you would have for a particular technology. So, definitely, battery it is not a zero emission technology, if you consider the lifetime. The other aspect uh, uh, from the um, the listener, is recycling. This will be crucially important. It will be a huge industry. And let me comment out from two perspectives. First, as economics. So, recycling will only be done if it’s econ- economically viable, and it will be economically viable for example, when the raw material cost goes up. So, people will be developing ways to recycle the battery, so that can be sold. So, if you think about the budget for recycling, it’s basically how much you get paid uh, for the material that you extract. On the engineering side, recycling is very interesting. When you think about recycling, you think about boiling it down to the fundamental building blocks. Maybe extracting cobalt again, that’s the most expensive element in a battery, but this is very challenging. So, one big opportunity could be, can we recycle battery without having to reduce it down to the original makeup of the battery? Can we for example, think about rejuvenating a battery, or regenerating a battery, as a way of recycle and reuse as opposed to deconstructing battery? And one- also they think about the safety of recycling. So, how do you take apart the battery safely and how do we design battery for recycling? So, can we design it with recycling in mind? So, the form factor already allows easy recycling. Great. Thank you so much. [NOISE] Here’s an interesting one [NOISE] that I [NOISE] pulled out. [NOISE] So, this participant wrote, California recently announced requiring [NOISE] solar panels on all new homes beginning in January first 2020. What are your thoughts on the effects of over generation, and potential curtail,- curtailment from this happening. Uh, the over generation [NOISE] is already there. So, as I showed in the duck curve, we have to ramp up, you know, half of the Australia grid in three hours, in California. This is how significant the problem is. So, um, this is only going to get worse. And storage is an obvious answer, but let me discuss it briefly what companies um, say for example, in California PG and E, and Southern California Edison are doing. So, there are a lot of legacy, power plant based on natural gas. And one way to do it is to simply power-up natural gas power plant. The hardware is already there, they’re not being fully utilized. So, that is the easy way. Battery based technology, or other technology is important as well, but currently, the investment needed to get that up and running basically, the capital expenditure needed, is enormous. So, looking forward, I think one thing that the policymaker should consider is, as we mandate more and more renewable electricity generation, we must also create policy to give incentive to more storage. And we’re seeing this already. For example, in California, we have this self-generation credit that um, give uh, several thousand dollars to homeowners to install uh, battery system um, at home. So, these are sort of complimentary policy that we’ll look at. I didn’t really mention the policy aspect today, but that’s another huge opportunity is to think about the role of policy in terms of shaping how renewable technology, included generation, and storage, will be deployed. Thank you so much. Now, last question that I think few people actually uh, asked this one. So, how can I get involved in, you know, this field especially if, you know, I have a lot of global experiences in power, but I don’t necessarily have startup money or [NOISE] startup experience? [NOISE] How could I, you know, as a professional get involved in, in energy storage? Yeah. I think this is also a very difficult question to answer. Um, I think you’ve taken the first step. Um, energy storage is extremely complex, and it is rapidly evolving, it changes by the month. There are new technologies emerging all the time. So, my recommendation would be to get caught up uh, by learning more about the technologies we have today and, continue to follow it. Um, there are many startups um, here in Silicon Valley and elsewhere, pursuing the development of energy technology. But I think as people who are interested in technology, you have to be able to assess it. So, what market is that technology addressing? What are the economical driving forces? What are the availability of scaling up? Right. You hear a lot of popular press and a lot of hype around, say battery technology, but I think one important skill to have is to be able to look at the technology and say, ”Okay, this is really going to have an impact at scale, or this is height.” And I think this is one of the things that um, you can try to learn for example, um, from our program here. Great. Well, thank you so much uh, everyone for joining us. We will be sending out a copy of this presentation uh, within a week, and thank you all. We hope you enjoy the rest of your day. Thank you.


  1. Interesting talk, albeit I doubt that solar + wind + battery storage solutions are the right answer for the needs of a carbon free industrial nation, at least not solely. We in Germany face tremendous problems due to our fast (and costly) ramp of renewable power generation without having meaningful power storage capacity. This causes all sorts of problems, not only for grid stability but also electricity prices – we now pay the highest energy prices in Europe. One absurd consequence is that we are now also more reliant on coal and gas than before. Phasing out nuclear energy entirely is harmful to our CO2 balance and grid stability. Phasing out Coal and Gas should have been the priority in my opinion, advancing Generation IV nuclear power instead. At least from current results, the French strategy after the oil crisis in the 1970th seems to be the better approach from a cost / benefit to counter climate change perspective.

  2. This could easily have been titled "The Next Big Opportunities for Lithium Batteries". In terms of "do it now", re-powering existing hydroelectric locations for pump-back is obvious, it's way cheaper than building pumped storage from scratch. For some unknown reason, Norway seems to be the only place that's figured this out?

    As a rule, the use of thermal storage is much cheaper than batteries, California's duck-curve could be solved with ice storage instead of electricity in the evening. In cold climates off peak thermal storage is also cheaper than batteries.

  3. Chemical energy storage is very promising….it allows for long term storage and regional distribution….it's not limited to hydrogen or methane either…I'm surprised other options weren't mentioned….my favorite being Ammonia currently. Many of us are proposing it as the next wave of the hydrogen economy.

  4. Nice informing video.Thanks!
    Where can i can the presentation from ?
    I'm a student and working on a project at the moment and would like to use some information.

  5. An interesting presentation in which the audience was expected to bring an electron microscope to see the small letters on the screen. Unfortunately, I forgot mine.

  6. 100% renewable energy worldwide isn’t just possible, it’s more cost-effective than existing system

  7. Thx for the informative lecture. But it would have been more enlightening, if there was an overview of the alternatives to storage for balancing the grid. Such as interconnecting grids worldwide or smart water heater tanks they or businesses that shift electric demand based on rates just to mention a few options. Since it was a little misleading, since the increased demand for storage could be offset by other alternatives.
    Or just focus on batteries witch was the main focus of the lecture.

  8. I think his dismissal of Pumped Hydro Storage is wrong. Batteries are great for frequency regulation but for large scale storage PHS – is cheap, has a very long calendar and cycle lifetime and can store energy for months with little loss. A report on 530,000 with the ability to store 100x + needed energy.

  9. You included a lot of emphasis on lithium technology. This relies on expensive chemicals and danger. The future for lithium batteries it's not the best choice. Your first category of energy storage was electromagnetic or ultracapacitors. They do not use expensive materials they're not dangerous when discharged and has the best long-term lifetime. The other energy storage categories are limited for their own respective technology. If you were to present the best possibility for future energy storage, would that be ultracapacitors?

  10. Really nice talk, but would have been easier to follow if he'd used a pointer. Also, consumer battery storage incentive is a complementary policy not a complimentary policy.

  11. Thanks for the lecture, it helped reinforce several ideas I've had about energy storage and in fact you have showed that we really missed an opportunity in 2015. But since we have been missing opportunities since about 1990, what else is new.
    All I will say is Windyday concept.

  12. According to the chart at 39:40, LIB will fall to $200 by 2025. But Tesla is already claiming it has much lower costs — near $100. As I understood it that was the pack cost.

  13. Pumped hydro for daily energy storage. For monthly or seasonal energy storage, produce hydrogen fuel gas by electrolysis of water. The cheapest back-up power is provided from biofuels.

  14. Wind, storage and back-up system designer for 100% renewable energy grids and microgrids with 24/7/52 power-on-demand!

  15. Put load frequency control on the hydro electric dams and store the water for future use. This would allow the use of the American infrastructure of dams that are government owned to produce more than just generating electricity.

  16. We need to start thinking more outside the box. Batteries aren't the only way to store energy. Solar Power Tower plants store solar energy as heat in molten salt tanks. This is 10 times less expensive than batteries.

    Also, fourth generation nuclear is looking a lot like a renewable. With these reactors, we have enough fuel to last hundreds of years. When that runs out, fusion reactors will power the planet for millions of years. And by the way, the best way to get rid of nuclear weapons – use them to make electricity. In fact, up to 10% of U.S. electricity today comes from decommissioned warheads.

    Solar and nuclear compliment each other. Nuclear is great at supplying constant baseline power. Solar output roughly matches our peak power demands during the day. Yes, we'll also need some batteries, mostly in sub-stations, with perhaps 2-4 hours of storage to help level out peaks. But this idea that we'll have enough batteries to store days worth of solar power when it rains, when you run those numbers, it just seems insane to me.

  17. Storage is the challenge and the criteria of success are: economically viable, operates silently, emissions-free, no moving parts, long lifespan (decades.)
    Of course capable of storing solar and wind for days and weeks, absorbing over-generation, shaving peak demand locally.
    Right now, that's carbon-nickel ("lithium ion") and in the future, maybe liquid metal (Chem C or a Magnesium/Salt/Antimony.)
    Tesla: The problems with battery technology "today" are rare raw materials and complex/expensive/dirty processes limited to about $100/kWh and short (10 year) lifespans.
    Ambri: A purpose-built system grid storage technology using abundant raw materials achieving $75/kWh.

  18. Energy, like clean air and water, like education and the Internet, must be provided by publicly owned and operated utilities, not corporate business operations.
    These "human rights" of technology must be decoupled from the profit motive and capitalism.
    A corporation cannot "own" the technology essential to quality of life to people at such a fundamental level, much less withhold it for a price to be paid.

  19. Pumped-storage hydro is by far the most economically feasible solution in large scale. It has not years not decades but over hundred years, the most mature engineering experience, no rare metal/elements invovled, simply steel and concrete based infrastructure building, and the MOST important factor is, steel/concrete lasts long long time, probably 100 or even 120 years life time, and need low level maintenance expense

  20. the challenge is do power companies, citezens and govenrment have enough money for uniform mini grids/off grid solar, micro wind, bio gas, teg's, and super caps. which i think there is. how is it an ecomnical blow to power companies when they are just trying to collect money and not improve the world.

  21. I think it is unwise to assume that lithium ion battery technology will continue to dramatically drop in cost. Fortunately there are other battery technologies like liquid metal that show great potential for storing large amounts of energy at low cost. The weight of the batteries used for grid storage is irrelevant because they would be stationary.

  22. Pack costs are below that showed in bar graphics, end of 2018 price was around 187$/kWh ( BNEF source) for automotive industry , minerals has decrease their price too: Cobalt price was 60% down and lithium 40% down

  23. This presentation is from late 2018. Why isn't Liquid Metal Battery storage discussed, even as a theoretical possibility?

  24. theres already some 25,000 GWH of storage in the tesla fleet, yes the older batteries probably arent the best or even the s/x but alot of the storage is now in model 3 batteries…the good new ones

    80kwh 3LR battery is 7000$ to replace or less than 100$/kwh

    at .1$/kwh delta which you can beat in quite a few places the car can make 8$/day sitting still or 3k/yr
    the battery is good for 1700 cycles or 4.7 full years of use and it would pay off in roughly 2 years…with 60% of its lifespan left

    realistically most people would be HAPPY to use 90% SOC to 40% SOC daily to help power the grid….gives the 4$/day uses 1/2 of a lifecycle, helps get time value back out of the car, helps save emissions etc etc etc

    if you already have solar why on earth cant the car act as a powerwall?
    doesnt chademo/ccs support backfeeding?

    you're never going to find a cheaper battery than the one you already own sitting in your car wasting away to time (devaluation)

  25. theres to be some 700,000 teslas in US by end of 2019? half of those would be the economical 3 with new longer lasting battery pack 1700 cycles…
    avg pack size of what 60kwh? so some 350,000*60kwh 21GWH.

    carbon free electricity is likely pointless. on cloudy days what are you going to do?
    you'd have to overbuild your battery so large to cover yearly trends (winter, storms etc) it would be uneconomical, 1 single powerplant kept operational across months may be able to provide enough power to a massive centralized battery during winter months to prevent huge amounts of batteries having to be installed to store energy from the summer months

    unless you just plan to run nuclear and hydro harder in the winter which I guess you could do

  26. the reference to the battery storage and use in Australia at 15:18 growing by 10% he forget to tell you that Victoria shut down one power station That is why the 10% of battery growth

  27. These advanced electricity storage technologies are only necessary BECAUSE we are choosing to use unreliable wind and solar. Therefore when we determine the cost of using wind and solar at greater than 10% of our needs, we need to factor in storage like giant battery farms. I would be curious to see a cost benefit analysis of 50% renewable vs. conventional, taking into account utility grid storage. I think conventional will still be hard to beat.

  28. One of the biggest opportunities, rarely discussed, is addressing the demand side of the duck curve. Policies that encourage real-time pricing of electricity that provides an incentive to use electricity at different times can be quite effective. A great deal of demand on the grid is for thermal management, heating and cooling, that can have many hours of latency. Simply heating or cooling your home, or charging your EV, at different times of the day or week can really help smooth out the duck curve.

  29. Why should a utility pay operators for renewable electricity they don't use? Probably some government mandate. But those adding renewable capacity should be made to consider whether they can always sell their energy. Otherwise they may end up adding capacity that doesn't make economic sense. And, that would also give them the incentive to explore storage. If they can get paid either way, why would they bother with storage? I guess the higher evening prices are a factor there. But the totality of the situation doesn't justify a utility paying for peak power it doesn't use.

  30. More people need to get involved with Energy Storage…This should have happened decades ago. I believe research at the University level and investment strategies for private firms needs to happen. Any research, of course, will have to be long term in thinking, I still believe that there are alternative sources of energy that have yet to be discovered. I think this is a more serious problem than is given priority today. What do you all think?

  31. Install PV near use, for example rooftops and parking lots (where the PV also shade the parked car).
    Install storage nearby, to reduce energy transport.
    Use existing EV fleet as extra storage (sink/source).

  32. The chart at 5:50 shows energy storage in GW (giga-watts) — which is a unit of power. Energy storage is measured in watt-hours (or giga-watt-hours, or joules, or whatever). Maybe it's a typo? (Although he talks about giga-watts of storage as well.)

  33. If you add the externalized enviromental costs, tax subsidies and asset protection costs by state militaries to basic hydocarbon fuel costs you'd find that renewables are easily competitive.

  34. I really with he talked more about future of PHS, and specifics of new technologies, like LTO (Lithium Titanium Oxide), Vanadium redox flow batteries and Zinc-bromide flow batteries. They appear to have extremely good levelized cost of energy.

    I am also interested in opinions on what is the best seasonal storage.

    And another aspect: how about building much bigger grid system, maybe even planet scale systems, so less storage is needed, as fluctuations in loads and in wind and solar are smoothed out a lot.

  35. One of my favourite energy storage systems is CAES by Highview Power here in the UK; on its own or along side a battery system it is very scalable and uses existing technology/parts.

  36. There are new technologies coming up… please take a look at New technology "beyond Lithium-Ion" based on Thermochemical storage.

Leave a Reply

Your email address will not be published. Required fields are marked *