Data for China🇨🇳

[paopao-mtf]

When did this happen?

2023-05-13

What zones are affected?

CN

What is the problem?

We would like to add data from China.

According to the Chinese government's report, we know that electricity is supplied by two main companies:

State Grid of China Ltd. China Southern Power Grid Co. We have some national wind energy statistics at the following link:https://sgnec.sgcc.com.cn/atlas/windResourceDistribution The following link has some national solar statistics: https://sgnec.sgcc.com.cn/atlas/solarEnergy It would be great to find some real time generation sources, or more data and statistics.

[paopao-mtf]

https://sgnec.sgcc.com.cn/renewableEnergy/developmentSituation Here's the nation's power structure

[paopao-mtf]

https://data.stats.gov.cn/easyquery.htm?cn=A01

This page allows you to find and call up monthly updates of the total national electricity generation and the respective generation of each energy source

[paopao-mtf]

http://cep.cec.org.cn/api/index.html Some data on national electricity prices can be found here

[paopao-mtf]

http://www.nengyuanyan.com/ This site also has some useful information

[jarek]

For the record, see previous issue with some links as well https://github.com/electricitymaps/electricitymaps-contrib/issues/1973

[qyearsley]

This would be awesome, I'd be really interested to see data for China on the map.

From before it sounds like the issue was lack of real-time data, but here it looks like there are monthly updates, would this be workable?

I'd be happy to help with parsers etc. if help is needed.

[celsomilne]

Likewise, would definitely help with a China parser!

[paopao-mtf]

Is the official attention to this issue?If possible, approximately when will China data be added?

[jarek]

Electricity Maps does not have data available. Electricity Maps is not aware of suitable data being available. Whoever controls this data would have to make it public and available.

Suitable data is: electricity generation values in MW or MWh broken down by source (hydroelectric, coal, nuclear, etc) in hourly granularity. Preferably this would be updated with a delay of 2 hours or less from actual generation.

[Nil-Cipher]

If data wasn't available in the hourly granularity, is monthly data not acceptable as an approximate? Something is better than nothing, right?

[qyearsley]

I just happened to be thinking about this today. Right now electricitymap's wiki page Technical requirements for a new parser says "Have a granularity of hourly or higher". As far as I understand, supporting data sources with low granularity (monthly etc.) would require some non-trivial updates to the UI -- but this might be worth considering for work, especially as some countries may be unlikely to provide high-granularity data in the next few years.

Meanwhile, I wanted to summarize and expand on the situation and available data for China as mentioned above by @paopao-mtf: There are two main "grid" state-owned enterprises that manage transmission, distribution, etc.:

• China Southern Power Grid (CSP, 中国南方电网), covering covering five southern provinces: Guangdong, Guangxi, Guizhou, Yunnan and Hainan.
• State Grid Corporation of China (SGGC, 国家电网公司), the rest of the provinces.

It's worth noting that the Chinese and English language versions of the websites for these two companies are quite different, obviously designed for different audiences. I linked the English versions.

Meanwhile, there are many other companies that manage power generation, including a few that may be called the "big 5" (SEIC, SPIC, Huaneng, Datang, Huadian).

Sources:

• https://energyiceberg.com/state-owned-power-utilities/
• https://en.wikipedia.org/wiki/State_Grid_Corporation_of_China
• https://en.wikipedia.org/wiki/China_Southern_Power_Grid

Meanwhile, the National Bureau of Statistics publishes a limited of data about lots of different things, including power generation, with monthly granularity, (not broken down by grid area). It also provides fossil fuel power generation as a single number that's not broken down by type (just called thermal power), e.g.: "Output of Thermal Power, Current Period(100 million kwh) 4712.5". Other sources listed area wind, solar, nuclear, and hydro. The interface isn't super great, you have to click through to "energy" and "output of energy products": https://data.stats.gov.cn/english/easyquery.htm

Then there's http://www.nengyuanyan.com/, which is apparently a "energy data sharing platform". Access to data is apparently not free. Includes stats about other energy resources, not just electricity.

[jarek]

Thanks for the data and links!

As far as I understand, supporting data sources with low granularity (monthly etc.) would require some non-trivial updates to the UI -- but this might be worth considering for work, especially as some countries may be unlikely to provide high-granularity data in the next few years.

The defining issue with exclusively monthly data is it doesn't give Electricity Maps a good idea of how carbon intensity changes in sub-monthly intervals. For example, if there's a lot of solar production, does demand peak during the solar production (e.g. because industry has been incentivized to use it then, or because it's stored in batteries or pumped hydro), or does demand continue into the evening causing a duck curve and requiring gas peakers? Or for another example, wind production might vary greatly throughout the month, but it'll be flattened out in the monthly average. Of conventional sources, are daily peaks met with fossil fuels, or hydro production?

As far as I know, in regions without hourly production data that Electricity Maps has integrated so far, at least they had hourly demand data, so there could be some estimates of demand vs likely solar production.

Without either demand or production data in sufficient granularity, it's just an unchanging colour on the map, and while there are some of those now (e.g. Indonesia?), as I understand it Electricity Maps would really like to be able to do estimates and predictions - more of an Electricity Carbon Data than an Electricity Maps.

[qyearsley]

Thanks @jarek , that makes a lot of sense -- upon an initial look at electricitymaps the first thing one sees is the overall map with a snapshot of carbon intensity for many regions, but if the goal was just to get a world map with average carbon intensity, that would be relatively easy. Specific time based estimates and predictions are more complicated (and valuable).

It's a shame that [I believe] we're unlikely to get high-granularity data for China. There's a lot going on in China, including some development of new renewable energy projects, but the state grid companies and bureau of statistics may not be interested in publishing high-granularity data unless they think it's worth it (which may be possible at some point?)

[paopao-mtf]

Thanks.I have found a document on the official website of the State Council of the People’s Republic of China. It is [PDF] Typical Power Load Curves of Provincial Power Grids - Gov.cn.Here is the link to the file(http://big5.www.gov.cn/gate/big5/www.gov.cn/zhengce/zhengceku/2020-12/03/5566580/files/eaaa93782e514543861bdcd434e86666.pdf) . This document can be used to estimate the hourly power grid load in various parts of mainland China during the day, and then calculate based on known energy structure information and reflect the results on the power map. Note that the unit here is MW. And the data here is divided into ordinary dates and holidays.I think the file may help us.Besides, it's better to have one than not to have one.

[VIKTORVAV99]

I think it would still be too many unknowns to create a good estimation model that can actually be used in production environments.

Estimating China is something that would have a huge impact on the carbon footprint on a very wide variety of products, companies and services which means we have to be able to at least validate our method to calculate/estimate these emission factors.

If we estimate every step of the process and let one estimation guide another there is a large risk we will introduce cascading errors and won't be able to trust our own data.

So at the very least we need to get consumption data live, and preferable production data in hourly or better values (even if they are delayed). It's probably possible to use daily values as well if we have hourly or better consumption but accuracy will go down the more we are estimating.

With all that said I'd love to get data for China but I don't think it's possible with the current data.

[paopao-mtf]

All right, I got it.Thank you for your reply, it really makes sense.