4 ms·

I know next to nothing about Solar so pardon my ignorance, but what do you suppose the solar industry is looking for from the software industry?

by jgh 10y ago

jonsen 10y ago
Storage technology ;)
diafygi 10y ago
Thanks for the comment! Gotcha, yes, it does take quite a bit of understanding of the dynamics and incentives of the industry to see how important software is to the industry. Here's a rundown of the problems that software can solve. Front End and User Experience: Customer acquisition costs in the U.S. for distributed solar is extremely high. Think of cleantech sales like the insurance industry 20 years ago. Most of the sales are human-to-human, and take a lot of time and legwork. We need to follow the path of the insurance industry, and start automating the initial quote and feasibility analysis. A consumer interested in solar or batteries or thermostats or whatever, should be able to get a savings quote instantly and buy the product online without having to talk to anyone. It's a complex problem (you have to combine energy pricing data inputs, easy user experience, complex data science, and dynamic pricing calculations), but the insurance industry has solved it, so we should, too. Whenever you go to Home Depot, you should be able to get a free energy audit with recommendations for what energy products would save you what. Software like this is how you scale energy efficiency and distributed energy resources. Data Science and High Scalability: Intermittent power from solar and wind will become a huge chunk of the grid's power. What happens when the sun doesn't shine or the wind doesn't blow? How do we keep the lights on? Storage? Demand response? It's going to be an extremely complex operation that will require much more complex software to be able to optimize deploying storage and demand response. If you want to do software in machine learning and NP optimization, grid operations is a place that needs it yesterday. IoT and Security: All of the new distributed energy resources (solar, storage, demand response, HVAC) are internet connected. The firmware and cloud software that run these resources are now a critical part of grid operations. If there's anything we've learned on HN, IoT software is complete shit when it comes to security. Beefing up the software security of all these resources is critical to keeping the lights on. That's all software and efficient protocols. The next Google will be an energy company. Edit: Shameless plug. If you're in the bay area and interested in getting involved in the cleantech software sector, check out the Powerhouse solar startup incubator (https://powerhouse.solar/ https://powerhouse.solar/) and my energy events calendar (https://bayareaenergyevents.com/ https://bayareaenergyevents.com/). Just start showing up!
- cderwin 10y ago
  > If there's anything we've learned on HN, IoT software is complete shit when it comes to security. Yes, but I think most people here will till you it's not because of a lack of engineering talent. Rather security takes time and a lot of money (in particular security engineers) and IoT companies would rather do without for the most part. And even if you do pay the price it doesn't always work out. It's not an engineering or talent problem, it's a management and P/L problem.
  - diafygi 10y ago
    True, however since the distributed energy resource (DER) IoT (inverters, thermostats, etc.) is more critical to grid infrastructure than your WiFi router, there's a bit more incentive to start doing more security testing on these deployed devices. Consumers won't pay for the better security, but grids likely will. Maybe it will take a huge hack to push the industry into requiring better DER security, or maybe an ambitious security firm could hop on this early and control the market. Either way, eventually we're going to end up with a set of security requirements for DERs that need to be tested and certified. Seems like a good opportunity for some ambitious tech cybersecurity startups. DERs are a really new sector of the energy industry, so there aren't any entrenched security firms (like there are in the wholesale grid). It's an open field for ambitious entrepreneurs.
- kozikow 10y ago
  Are solar panel inspections via drone a thing? E.g. find cracks in panels, locate all panels in the area, detect bad modules, etc. Shameless plug: We (tensorflight.com) do computer vision on drone imagery and heard a customer asking about solar panels. We are in Mountain View, so please get in touch if you have use case in mind.
  - diafygi 10y ago
    Not really. You can remotely tell if a system is broken based on the power it producing, and you usually clean the panels at least every few years. So any cracks would be discovered relatively quickly, but they are pretty rare anyway. The glass they use to protect the panels is super strong, and designed to withstand large hail. The only time I've seen broken panels is when someone hits a baseball into one. However, drone imagery is heavily used in the solar industry for roof inspection. For rooftop solar, you need to inspect the roof to make sure the panels will fit and won't be obstructed. Google maps can give you a ballpark, but closer imagery is needed. There's a few companies doing this (Solar Census, Google Sunroof, Sun Number, etc.), so might check them out. FYI, the coolest cleantech application I've see for areal imagery is Ceres Imaging, who flies over crops to detect where you need to water. Something like a 50%+ water savings because you can target water to where it's needed. Crazy effective.
rogermedia 10y ago
From the little experience I have with solar what I've noticed is that its mostly a hardware issue. Storage is the biggest cost now. Which is why you see things like the gigafactory. We've got the ability to produce electricity from wind and solar down pretty good. There are efficiencies to improve but storage is the main block. My experience with batteries is that lithium ion on a large scale is still too expensive and there are quality issues. For that reason I use lead acid batteries in my home. They are a tenth the cost and a proven and reliable technology. Meaning if I use them correctly they can last for five to seven years. I know people who use lithium ion in their home and either they built them or spent a lot of money on them. In most cases the system has not lived up to its promises and from a cost perspective still isn't quite there. If I were to upgrade now it would be to AGM batteries which are similar to lead acid but easier to manage. If I were smart enough not to get electrocuted I'd be playing around with this stuff more. Not sure how much software is in that though. One area where software would be helpful is potentially with the distributed grid stuff. You have a large set of panels in a neighborhood that create power shared to everyone. It seems there's a lot to keep track of with something like that. Solar city has a pretty good lock on things until the consumer understands it's a terrible deal for them. They are leasing the solar panels to you and bolting them into your roof. it would be like leasing your water heater except 100x more expensive. We went with a local installer for our system and saved money and had a great experience. the vc people don't have the potential of making tons of money like in solar city's case.

I know next to nothing about Solar so pardon my ignorance, but what do you suppose the solar industry is looking for from the software industry?

by jgh 10y ago

I know next to nothing about Solar so pardon my ignorance, but what do you suppose the solar industry is looking for from the software industry?

jonsen 10y ago

Storage technology ;)

diafygi 10y ago

Thanks for the comment! Gotcha, yes, it does take quite a bit of understanding of the dynamics and incentives of the industry to see how important software is to the industry. Here's a rundown of the problems that software can solve. Front End and User Experience: Customer acquisition costs in the U.S. for distributed solar is extremely high. Think of cleantech sales like the insurance industry 20 years ago. Most of the sales are human-to-human, and take a lot of time and legwork. We need to follow the path of the insurance industry, and start automating the initial quote and feasibility analysis. A consumer interested in solar or batteries or thermostats or whatever, should be able to get a savings quote instantly and buy the product online without having to talk to anyone. It's a complex problem (you have to combine energy pricing data inputs, easy user experience, complex data science, and dynamic pricing calculations), but the insurance industry has solved it, so we should, too. Whenever you go to Home Depot, you should be able to get a free energy audit with recommendations for what energy products would save you what. Software like this is how you scale energy efficiency and distributed energy resources. Data Science and High Scalability: Intermittent power from solar and wind will become a huge chunk of the grid's power. What happens when the sun doesn't shine or the wind doesn't blow? How do we keep the lights on? Storage? Demand response? It's going to be an extremely complex operation that will require much more complex software to be able to optimize deploying storage and demand response. If you want to do software in machine learning and NP optimization, grid operations is a place that needs it yesterday. IoT and Security: All of the new distributed energy resources (solar, storage, demand response, HVAC) are internet connected. The firmware and cloud software that run these resources are now a critical part of grid operations. If there's anything we've learned on HN, IoT software is complete shit when it comes to security. Beefing up the software security of all these resources is critical to keeping the lights on. That's all software and efficient protocols. The next Google will be an energy company. Edit: Shameless plug. If you're in the bay area and interested in getting involved in the cleantech software sector, check out the Powerhouse solar startup incubator (https://powerhouse.solar/ https://powerhouse.solar/) and my energy events calendar (https://bayareaenergyevents.com/ https://bayareaenergyevents.com/). Just start showing up!

cderwin 10y ago
> If there's anything we've learned on HN, IoT software is complete shit when it comes to security. Yes, but I think most people here will till you it's not because of a lack of engineering talent. Rather security takes time and a lot of money (in particular security engineers) and IoT companies would rather do without for the most part. And even if you do pay the price it doesn't always work out. It's not an engineering or talent problem, it's a management and P/L problem.
- diafygi 10y ago
  True, however since the distributed energy resource (DER) IoT (inverters, thermostats, etc.) is more critical to grid infrastructure than your WiFi router, there's a bit more incentive to start doing more security testing on these deployed devices. Consumers won't pay for the better security, but grids likely will. Maybe it will take a huge hack to push the industry into requiring better DER security, or maybe an ambitious security firm could hop on this early and control the market. Either way, eventually we're going to end up with a set of security requirements for DERs that need to be tested and certified. Seems like a good opportunity for some ambitious tech cybersecurity startups. DERs are a really new sector of the energy industry, so there aren't any entrenched security firms (like there are in the wholesale grid). It's an open field for ambitious entrepreneurs.
kozikow 10y ago
Are solar panel inspections via drone a thing? E.g. find cracks in panels, locate all panels in the area, detect bad modules, etc. Shameless plug: We (tensorflight.com) do computer vision on drone imagery and heard a customer asking about solar panels. We are in Mountain View, so please get in touch if you have use case in mind.
- diafygi 10y ago
  Not really. You can remotely tell if a system is broken based on the power it producing, and you usually clean the panels at least every few years. So any cracks would be discovered relatively quickly, but they are pretty rare anyway. The glass they use to protect the panels is super strong, and designed to withstand large hail. The only time I've seen broken panels is when someone hits a baseball into one. However, drone imagery is heavily used in the solar industry for roof inspection. For rooftop solar, you need to inspect the roof to make sure the panels will fit and won't be obstructed. Google maps can give you a ballpark, but closer imagery is needed. There's a few companies doing this (Solar Census, Google Sunroof, Sun Number, etc.), so might check them out. FYI, the coolest cleantech application I've see for areal imagery is Ceres Imaging, who flies over crops to detect where you need to water. Something like a 50%+ water savings because you can target water to where it's needed. Crazy effective.

rogermedia 10y ago

From the little experience I have with solar what I've noticed is that its mostly a hardware issue. Storage is the biggest cost now. Which is why you see things like the gigafactory. We've got the ability to produce electricity from wind and solar down pretty good. There are efficiencies to improve but storage is the main block. My experience with batteries is that lithium ion on a large scale is still too expensive and there are quality issues. For that reason I use lead acid batteries in my home. They are a tenth the cost and a proven and reliable technology. Meaning if I use them correctly they can last for five to seven years. I know people who use lithium ion in their home and either they built them or spent a lot of money on them. In most cases the system has not lived up to its promises and from a cost perspective still isn't quite there. If I were to upgrade now it would be to AGM batteries which are similar to lead acid but easier to manage. If I were smart enough not to get electrocuted I'd be playing around with this stuff more. Not sure how much software is in that though. One area where software would be helpful is potentially with the distributed grid stuff. You have a large set of panels in a neighborhood that create power shared to everyone. It seems there's a lot to keep track of with something like that. Solar city has a pretty good lock on things until the consumer understands it's a terrible deal for them. They are leasing the solar panels to you and bolting them into your roof. it would be like leasing your water heater except 100x more expensive. We went with a local installer for our system and saved money and had a great experience. the vc people don't have the potential of making tons of money like in solar city's case.