研究 04 · 情境模拟Research 04 · Scenario simulation
气候能源转型模拟器Climate energy transition simulator
一户人家该不该把燃煤炉换掉？换成天然气，还是电，还是热泵？这是个应用数学建模项目，做成了可以交互的网站，分中国政策情境模拟和全球家庭能源顾问两个版本。Should a household give up its coal stove, and for gas, for electricity, or for a heat pump? This applied modeling project answers that as an interactive site, in two versions: a policy scenario simulator for China and a household energy adviser for anywhere.
中国版：把政策放进一户人家的账里China version: policy, as one household's account
变量取家庭收入、储蓄、住房面积、供暖支出和当前的技术选择。模型逐年推进，每一年的状态由上一年的状态与当年的选择决定。The variables are household income, savings, floor area, heating expenditure and the technology currently in use. The model steps forward year by year, each year's state following from the previous one and from the choice made.
煤改气、煤改电和热泵各走一条路径。改造成本、运行费用、碳排放、政策合规度都按分段函数算，补贴按档位跳变，收入和气温则带随机成分。模型最后给出的，是能源负担率随年份的变化，以及补贴退出之后这户人家会不会撑不住。Coal-to-gas, coal-to-electricity and heat pumps each follow their own path. Retrofit cost, running cost, carbon emissions and policy compliance are computed with piecewise functions, subsidies jump between tiers, and income and temperature carry a random component. What comes out is the energy burden over time, and whether the household can still carry it once the subsidy ends.
能源负担率是其中最要紧的一个数。设备装好而电费难以承担，改造便等于没有完成。补贴退出的那一年，这个比率往往跳一下，模型把这一跳明确算出来。The energy burden is the number that matters most. A household that installs the equipment and then cannot pay the electricity bill has not really made the switch. In the year a subsidy ends the ratio tends to jump, and the model reports that jump explicitly.
全球版：先估需求，再排方案Global version: estimate demand, then rank the options
国别一换，补贴政策随之不同，可依凭的只剩更基础的量。全球版从采暖度日和制冷度日入手，配上能源账单、设备效率和当地能源价格，估算一户人家的冷热需求。Move to another country and the subsidy rules change, so only the more basic quantities survive. The global version starts from heating and cooling degree days, adds the energy bill, equipment efficiency and local energy prices, and estimates the household's heating and cooling demand.
可行的方案先筛一遍，剩下的按四个维度打分加权，再排序：可负担性、气候适应性、环境表现、实施便利性。Infeasible options are filtered out first; the rest are scored on four dimensions, weighted and ranked: affordability, climate suitability, environmental performance and ease of implementation.
热泵在温和气候里通常排在前面，到了极冷地区，效率掉下来，排序就可能翻过来。这类翻转正是模型要抓的地方。Heat pumps usually come out on top in mild climates. Where winters are severe their efficiency falls and the ranking may reverse, and reversals of that kind are exactly what the model is built to catch.
规则写在明面上The rules are written down
数据缺失、极端气候、基础设施限制，模型对每种情形都设了明确规则，不靠临时判断。由此每一项结果都能追溯，也能复算：换一组输入，结果为何改变，看得出来。Missing data, extreme climates and infrastructure limits are each handled by an explicit rule rather than a judgement call made on the spot. The result is that every output can be traced and recomputed: change an input and you can see why the answer moved.
做成网站，是因为这类问题的答案与具体那户人家绑得很紧。参数一改，结论就可能反过来；与其给一条平均意义上的建议，不如让人把自己的数字填进去。It became a website because the answer is bound tightly to the particular household. Change the parameters and the conclusion can flip, so it is better to let people enter their own numbers than to hand out advice that holds only on average.