Utafiti huu umeunda modeli ya uboreshaji wa vipimo vingi vya muda kwa ajili ya kuendesha kwa pamoja katika masoko ya umeme na kaboni bustani ya source-grid-load-storage inayounganisha upepo, jua, turbine ya gesi yenye carbon capture, battery storage, flexible electric loads na vifaa vya power-to-gas. Modeli inaunganisha day-ahead electricity market, intraday electricity market, real-time electricity market, ancillary services market na carbon quota market katika muundo mmoja wa maamuzi.

Watafiti wamegawanya uendeshaji katika tabaka tatu zinazofuatana. Katika hatua ya day-ahead, mpango wa saa 24 huundwa kwa time steps za saa 1; uamuzi wa kuwasha/kuzima turbine ya gesi, ununuzi na uuzaji wa umeme, reserve capacity na shiftable loads huamuliwa. Katika hatua ya intraday, moving window ya saa 2 huboreshwa upya kwa resolution ya dakika 15; forecast deviations, reducible loads na intraday electricity transactions husahihishwa. Katika hatua ya real-time, moving window ya dakika 15 hutatuliwa kwa steps za dakika 5, na ununuzi-uuzaji wa umeme, battery, carbon capture system na power-to-gas equipment hurekebishwa kwa madhumuni ya balancing.

Katika base scenario, total day-ahead operating cost ilihesabiwa kuwa 88,51×10³ yuan, intraday cost 75,53×10³ yuan na real-time cost 69,20×10³ yuan. Kutoka day-ahead hadi real-time, cost ilipungua kwa takriban %21,82. Kupungua huku kulihusishwa na kusasishwa kwa forecasts kadiri muda unavyokaribia na kusahihishwa kwa operating decisions kwa tabaka.

Comparison results zinaonyesha kwamba kujiondoa katika ancillary services market huongeza real-time cost kwa %2,89. Carbon market ilipoondolewa, real-time cost ilipanda kutoka 69,20×10³ yuan hadi 95,29×10³ yuan; ongezeko lilikuwa takriban %37,70. Park iliposhiriki katika day-ahead electricity market pekee bila intraday na real-time correction, real-time cost iliongezeka hadi 128,62×10³ yuan; takriban %85,87 juu ya base scenario.

Katika comparison scenario ambapo day-ahead plan pekee ilitumika, forecast deviations hazikuweza kutatuliwa kwa internal resources; katika intraday na real-time stages, wind-solar curtailment na unmet load zilitokea. Katika real-time table, renewable energy curtailment cost ilihesabiwa kuwa 21,23×10³ yuan na unmet load cost 8,28×10³ yuan.

Utafiti pia unaonyesha kwamba carbon quota hubadilika pamoja na electricity operating decisions. Katika base scenario, calculated carbon emissions za park ni 79.025,12 kg katika day-ahead stage, 47.959,98 kg katika intraday stage na 52.923,00 kg katika real-time stage. Katika hatua hizo hizo, excess quota inayodhaniwa kuwa inaweza kuuzwa katika carbon market ni 314.232,73 kg, 319.658,11 kg na 312.130,01 kg mtawalia.

Matokeo haya hayakutokana na industrial park halisi, bali na numerical example iliyoundwa na watafiti. Modeli inaratibu components zote ndani ya park karibu na power balance moja; hata hivyo, haitatui kwa undani voltage, line capacity, power loss na security limits za electrical grid. Kwa hiyo, utafiti unatoa economic energy-management model, si detailed distribution-grid validation.

Umuhimu unaowezekana kwa Uturuki

Utafiti haukutumia electricity price, carbon price, industrial load, wind-solar generation au market rules za Uturuki. Hata hivyo, kuunganisha day-ahead planning, intraday deviation correction na real-time balancing katika energy-management system moja kunaweza kurekebishwa kimetodolojia kwa organized industrial zones, energy-intensive facilities, campuses, ports na mixed generation-storage facilities.

Ili kuhamishwa kwa Uturuki, free carbon quota, carbon price, emission factor ya electricity purchased from grid, ancillary-service payment structure na transmission-distribution charges zilizo specific kwa China katika utafiti lazima zibadilishwe kabisa. Aidha, local grid connection capacity, transformer na line limits, generation licenses, storage rules, market participation conditions na real facility measurements lazima ziongezwe kwenye modeli.

Cost ya 69,20×10³ yuan, %21,82 reduction au quota sales ya hundreds of tons katika utafiti si expected result kwa industrial park yoyote nchini Uturuki. Hizi ni simulation outputs tu zilizopatikana chini ya system, price series, free-quota mechanism na equipment capacities zilizoundwa na watafiti.

Tatizo kuu la utafiti ni nini?

Katika energy parks zinazotumia kiwango kikubwa cha wind na solar generation, generation na consumption hazibadiliki kila mara katika mwelekeo mmoja. Solar generation inaweza kupanda mchana wakati electricity demand inafikia kilele jioni. Wind generation inaweza kuwa juu au chini ya forecast. Electricity price inaweza pia kubadilika katika intraday na real-time markets baada ya day-ahead planning.

Ikiwa system inaendeshwa kwa day-ahead forecast pekee, deviations zinazotokea katika saa zinazofuata zinaweza kusababisha expensive electricity purchase, renewable-energy curtailment au unmet load. Tofauti katika response times za resources kama battery, flexible load, gas turbine na P2G pia zinaweza kufanya single-time-scale optimization kutotosha.

Tatizo la pili ni kwamba electricity transactions na carbon transactions huendeshwa tofauti katika modeli nyingi. Hata hivyo, purchasing electricity from grid, operating gas turbine, energy consumption ya carbon capture na matumizi ya renewable generation hubadilisha net carbon position ya park. Carbon-quota price inaweza pia kubadilisha economic ordering ya electricity generation na purchase decisions.

Watafiti waliunganisha matatizo haya mawili na kutafuta jibu kwa swali hili: Masoko ya umeme na kaboni yanaposimamiwa katika common cost function, kwa day-ahead–intraday–real-time layers zinazofaa response speeds za equipment tofauti, cost, carbon position na operational flexibility ya park hubadilikaje?

Muundo wa source-grid-load-storage park

Kielelezo 2 katika ukurasa wa 6 wa utafiti kinaonyesha energy na information connections za park. Components kuu za park ni:

Park imechukuliwa kuwa price taker katika electricity market. Kwa maneno mengine, transactions za park hazibadilishi market price. Modeli huchagua generation, consumption, electricity buy-sell, reserve capacity na carbon-quota amount kwa kuangalia forecast electricity na carbon prices.

Electricity na carbon transactions zinaunganishwaje?

Kielelezo 1 katika ukurasa wa 4 wa utafiti kinaonyesha aina mbili za coupling. Coupling ya kwanza ni kwamba electricity na carbon decisions hufanywa pamoja. Katika traditional sequential method, electricity transactions hukamilishwa kwanza, kisha carbon quota huhesabiwa kutoka emissions zilizotengenezwa na transactions hizo. Katika method iliyopendekezwa, electricity buy-sell decision na carbon-quota decision huathiriana.

Coupling ya pili ni kati ya time scales. Electricity market huzalisha hourly day-ahead decisions, 15-minute intraday decisions na 5-minute real-time decisions. Katika utafiti, carbon price hugawanywa katika intervals hizo hizo bila kubadilishwa. Kwa hivyo, corresponding carbon cost huongezwa pia kwenye objective function katika kila electricity decision interval.

Mkabala huu unamaanisha kutumia carbon price ileile katika finer time slices. Kwa hiyo, high-frequency real carbon price haitabiriwi; carbon effect ya electricity schedule huhesabiwa upya mara nyingi zaidi.

Time scales tatu zinafanyaje kazi?

Baadhi ya maamuzi yaliyowekwa day-ahead huhamishwa kwenye hatua zinazofuata kama fixed data. Gas-turbine on-off status haiwezi kubadilishwa intraday au real-time; ni power yake na operating ratio ya carbon-capture equipment pekee zinarekebishwa kwa finer resolution.

Intraday model husogezwa mbele kila dakika 15 na saa 2 zinazofuata huhesabiwa upya. Real-time model husogezwa mbele kila dakika 5 na kutatua dakika 15 zilizo mbele yake. Muundo huu ni matumizi ya model predictive control katika energy market.

Wind na solar generation

Wind na photovoltaic generation zimewekewa kikomo zisizidi forecast available power:

\[ 0\leq P_t^{wt}\leq \theta_{\max}^{wt}P_0^{wt} \]

\[ 0\leq P_t^{pv}\leq \theta_{\max}^{pv}P_0^{pv} \]

\(P_t^{wt}\) na \(P_t^{pv}\) zinaonyesha generation katika time interval husika, \(P_0\) installed capacity, na \(\theta_{\max}\) maximum generation ratio inayotegemea forecast au availability.

Katika example park, wind installed capacity ni 50 MW na photovoltaic installed capacity ni 26,60 MW. Kielelezo 5–7 kinaonyesha kwamba kadiri time scale inavyofupishwa, generation na load forecasts zinakuwa more variable na detailed.

Gas turbine yenye carbon capture

Net electrical output ya gas turbine imehesabiwa kwa kuondoa electricity consumption ya carbon-capture system kutoka gross generation:

\[ P_t^{ccgt}=P_t^{gt}-P_t^{cc} \]

Consumption ya carbon-capture equipment ni jumla ya fixed consumption na variable consumption inayotegemea captured carbon-dioxide amount. System pia ina split ratio inayobainisha ni kiasi gani cha flue gas kinapitishwa kwenye capture equipment.

Net emission ya gas turbine inayobaki moja kwa moja kwenda atmosphere:

\[ E_{CO_2,t}^{ccgt} = \left(1-\beta^{cc}\theta_t^{cc}\right) \delta_{CO_2}^{gt}P_t^{gt}\Delta t \]

imemodeliwa hivi. \(\beta^{cc}\) inaonyesha capture efficiency, \(\theta_t^{cc}\) fraction ya flue gas inayoelekezwa kwa capture, na \(\delta_{CO_2}^{gt}\) emission factor ya gas turbine.

Katika example, nominal gas-turbine power ni 17,81 MW, nominal carbon-capture consumption power 1,61 MW, capture efficiency %90 na flue-gas split ratio iko katika range ya %15–%85.

Flexible load model

Load imegawanywa katika classes tatu. Base load haiwezi kubadilishwa kwa muda. Shiftable load huhamishwa hadi low-price hours huku total daily energy ikihifadhiwa. Reducible load hupunguzwa kwa muda kwa malipo kwa user.

Total load imetolewa kama:

\[ P_t^{load}=P_t^{cl}+P_t^{sl}+P_t^{rl} \]

Hapa \(P_t^{cl}\) ni base load, \(P_t^{sl}\) shiftable load na \(P_t^{rl}\) reducible load.

Shiftable load inasimamiwa kwa day-ahead price signal. Katika utafiti, saa ambazo load inaweza kuingizwa ni 00.00–08.00, na saa ambazo inaweza kuhamishwa nje ni 10.00–12.00 na 17.00–22.00.

Malipo ya user kwa reducible load ni:

\[ C^{rl}=\sum_t \alpha^e\lambda_t^e\Delta P_t^{rl} \]

\(\alpha^e=0{,}8\) ni incentive coefficient. Maximum load ni 50 MW, shiftable-load ratio limit %60 na reducible-load ratio limit %50.

Energy storage model

Energy balance ya battery imetolewa katika utafiti kama:

\[ E_t^{es} = E_{t-1}^{es} + P_t^{chr-es}\Delta t - \frac{P_t^{dis-es}\Delta t}{\eta_{dis}} \]

Battery haiwezi charge na discharge kwa wakati mmoja, energy level huhifadhiwa kati ya lower na upper limits, na energy ya mwisho wa siku huwekwa sawa na initial value.

Example battery ina power ya 15,32 MW, energy capacity ya 30,64 MWh, discharge efficiency %90 na full charge-discharge duration ya saa 2. Kwa kuwa hakuna separate charging efficiency iliyotolewa, equation inaonekana kuchukulia charging process kuwa lossless.

Power-to-gas conversion

Wakati renewable generation ni surplus, P2G equipment hutumia electricity kutengeneza gas:

\[ Q_t^{pg}=\eta_{pg}P_t^{pg} \]

Katika modeli, hydrogen iliyozalishwa hu-react na carbon dioxide kutoka carbon-capture system na kubadilishwa kuwa methane. Example P2G power ni 1,63 MW na conversion coefficient ni 0,15 m³/kWh.

Ingawa modeli ina material balance kwa methane production, haina intermediate hydrogen storage, electrolyzer startup time, compression energy au partial-load efficiency variation.

Carbon-trading cost

Day-ahead carbon transaction imehesabiwa kwa muundo huu:

\[ C_{da}^{CO_2} = \sum_t \lambda_t^{CO_2} \left[ \left( \delta_{CO_2}^{grid}P_{da,t}^{b,e} + \frac{E_{da,t}^{ccgt}}{\Delta t} \right) - \delta_{CO_2}^{free} \left( P_{da,t}^{load}+P_{da,t}^{pg} \right) \right]\Delta t \]

Bracket ya kwanza inawakilisha emissions zinazotokana na electricity purchased from grid na gas turbine, huku bracket ya pili ikiwakilisha free quota allocated kulingana na park load na P2G consumption. Ikiwa net value ni positive, quota purchase cost hutokea; ikiwa negative, revenue ya kuuza excess quota hutokea.

Katika utafiti, emission factor ya electricity purchased from grid ilitumika kuwa 0,531 kg/kWh, na free-quota coefficient 0,428 kg/kWh. Muundo huu wa allocation ya carbon quota ni mojawapo ya assumptions muhimu za example model.

Power balance ya park

Day-ahead power balance inategemea kusawazisha generation na inputs zote na consumption na outputs zote:

\[ P_t^{wt}+P_t^{pv}+P_t^{gt}+P_t^{dis-es}+P_t^{buy} = P_t^{load}+P_t^{pg}+P_t^{chr-es}+P_t^{cc}+P_t^{sell} \]

Equation hii inahakikisha total energy balance ya park. Hata hivyo, haionyeshi electricity flow inapitia lines gani ndani ya park, ikiwa voltages zinabaki katika acceptable range, au short-circuit na thermal limits katika connection point.

Day-ahead objective function

Day-ahead cost ni jumla ya vipengele vifuatavyo:

• Maintenance costs za wind, solar, gas turbine, carbon capture, battery na P2G
• Gas-turbine start-stop costs
• Day-ahead electricity buy-sell cost
• Natural-gas purchase cost
• Carbon-quota cost au revenue
• Upward na downward reserve revenue kutoka ancillary services

Katika electricity purchase, si market price pekee iliyotumika, bali pia 0,14 yuan/kWh transmission-distribution, public fund na surcharge.

Intraday na real-time objective functions

Katika intraday stage, gas-turbine start-stop decision na ancillary-service bid haziamuliwi upya kwa sababu zimehamishwa kutoka day-ahead. Intraday electricity difference transactions na reducible-load incentives huongezwa katika hatua hii.

Katika real-time stage, electricity buy-sell calculation inajumuisha deviation kutoka day-ahead na intraday schedules. Battery, carbon-capture energy consumption na P2G power hutumiwa kwa steps za dakika 5 ili kuanzisha final balance.

Example system parameters

Comparison scenarios

Day-ahead results

Katika day-ahead base scenario, park load ilikidhiwa na wind, solar, gas turbine na market purchase. Market purchase ilifanywa katika saa za low electricity price; katika high-price period ya 18.00–22.00, gas turbine yenye carbon capture iliendeshwa na battery ikadischarge. Katika baadhi ya saa electricity iliuzwa kwenye market.

Negative ancillary-service na carbon costs zinamaanisha revenue. Katika base scenario, park ilihesabiwa kupokea 393.257,85 kg free quota, kutumia 79.025,12 kg quota kusawazisha emissions zake na kuuza remaining 314.232,73 kg quota.

Intraday results

Katika intraday stage, new 15-minute forecasts zilitumika, reducible loads zikawezeshwa na deviations zikarekebishwa kwa intraday electricity transactions. Electricity buy-sell item iligeuka kuwa net revenue ya −11,75×10³ yuan.

Katika base scenario, intraday free quota ni 367.618,09 kg, quota iliyotumika katika internal balancing ni 47.959,98 kg na sold quota ni 319.658,11 kg.

Katika Comparison 3 ambapo day-ahead plan pekee ilitumika, kwa sababu intraday market correction haikuwezekana, renewable-generation curtailment ilitokea katika baadhi ya vipindi na load loss katika vipindi vingine. Total cost ilipanda hadi 121,77×10³ yuan.

Real-time results

Katika real-time stage, battery, P2G na carbon-capture consumption zilirekebishwa kwa steps za dakika 5. Kielelezo 16 kinaonyesha kwamba battery ilicharge katika low-price periods na kudischarge katika high-price periods.

Katika real-time base scenario, free quota ilihesabiwa kuwa 365.053,01 kg, amount iliyotumika katika internal balancing 52.923,00 kg na excess quota inayoweza kuuzwa kwenye market 312.130,01 kg.

Cost comparison

Thamani katika table ni katika units za 10³ yuan. Economic deterioration kubwa zaidi ilionekana katika scenario ambako intraday na real-time electricity markets ziliondolewa. Sababu haikuwa expensive electricity pekee, bali pia renewable-energy curtailment na unmet-load penalties.

%21,82 cost reduction ilihesabiwaje?

Ratio katika summary ya utafiti inategemea calculation hii:

\[ \frac{88{,}51-69{,}20}{88{,}51}\times100 \approx21{,}82\% \]

Comparison hii ni kati ya day-ahead planned cost ya base scenario na real-time updated cost ya scenario hiyo hiyo. Kwa hiyo, ratio haipaswi kutafsiriwa kama “multi-time-scale model ni %21,82 cheaper kuliko model nyingine”. Kauli sahihi zaidi ni kwamba calculated daily cost ilipungua kwa %21,82 dhidi ya day-ahead forecast baada ya plan kutatuliwa upya kwa more up-to-date forecasts na finer adjustments.

Effect ya carbon market

Carbon market ilipoondolewa, negative carbon cost ya base scenario, yaani quota-sale revenue, ilitoweka. Wakati huo huo generation na electricity-purchase ordering zilibadilika na park ikawa dependent zaidi kwa electricity purchased from grid.

Sehemu muhimu ya kuongezeka kwa real-time cost kutoka 69,20 hadi 95,29×10³ yuan inatokana na kupotea kwa carbon revenue ya 21,61×10³ yuan. Kwa hiyo, utafiti unaweka example ambako carbon market si emission constraint pekee, bali pia major revenue item.

Hata hivyo, kuwa na hundreds of tons of excess quota katika stages zote tatu kunafanya economic advantage ya carbon trading kutegemea sana free-quota allocation. Jinsi result ingebadilika kwa quota ndogo au carbon price tofauti haikuchunguzwa.

Effect ya ancillary services

Ancillary services zilipoondolewa, cost increase ilibaki relatively limited. Real-time cost iliongezeka kutoka 69,20 hadi 71,20×10³ yuan, difference ikiwa %2,89.

Ingawa direct value ya ancillary-service revenue ni ndogo, ilibadilisha gas-turbine decision ya kushikilia controllable reserve capacity na hivyo kuleta indirect effects kwenye electricity buy-sell na natural-gas costs.

Effect ya multi-time-scale electricity market

Comparison 3 inawakilisha hali ambapo day-ahead schedule haiwezi kubadilishwa katika stages zinazofuata. Kwa sababu forecast errors hazikurekebishwa kwa intraday na real-time electricity transactions, internal flexible resources pekee hazikutosha.

Katika intraday stage, total cost ilipanda hadi 121,77×10³, na katika real-time hadi 128,62×10³ yuan. Katika real-time result, renewable curtailment cost ni 21,23×10³ yuan na unmet load cost ni 8,28×10³ yuan.

Hii ni finding yenye nguvu zaidi ya modeli: wakati forecast wind, solar na load values zinabadilika, kufunga trading plan siku moja mapema pekee kunaweza kusababisha serious balancing problems katika park yenye high renewable generation.

Scientific message ya figures

Nguvu za utafiti

• Electricity na carbon transactions zimeunganishwa katika optimization structure moja.
• Day-ahead, intraday na real-time decisions zimetenganishwa wazi.
• Equipment zimesambazwa kwenye control layers kulingana na response times tofauti.
• Wind, solar, gas turbine, carbon capture, battery, P2G na demand response zote ziko katika modeli moja.
• Effects za ancillary services, carbon market na multi-stage electricity markets zimechunguzwa kwa comparisons tofauti.
• Carbon quota imetumika si kama indicator inayohesabiwa baadaye tu, bali kama direct economic decision variable.
• Layered solution process inayotegemea YALMIP na Gurobi imeelezwa.
• Rolling-window optimization imetumika katika intraday na real-time stages.
• Separate penalty costs zimefafanuliwa kwa renewable curtailment na unmet load.
• Utafiti umeshiriki detailed cost components na carbon-quota flows.

Vikwazo vya utafiti

• Utafiti ni preprint ambayo haijapitiwa na rika.
• Hakuna real industrial-park measurement iliyotumika.
• Parameters za kutengeneza forecast data hazijaelezwa.
• Single forecast path imetatuliwa badala ya stochastic au robust optimization.
• Model solution time na optimality gap hazijatolewa.
• Electrical grid haijamodeliwa kwa physical power flow.
• Hakuna voltage, line current, transformer capacity, loss na security constraints.
• Battery charging efficiency na cycle aging hazijamodeliwa wazi.
• Carbon-capture na P2G systems zimerahisishwa kwa fixed coefficients.
• High-frequency market yenye carbon price inayobadilika kwa muda haijamodeliwa.
• Hakuna sensitivity analysis kwa free quota na carbon price.
• Effect ya demand response kwa production, comfort au industrial process haijatathminiwa.
• Natural-gas infrastructure na methane-storage capacity hazijamodeliwa kwa undani.
• Hakuna strategic price-making behavior katika electricity na carbon markets.
• Baadhi ya sentensi zinazoeleza cost tables zinasema kinyume cha table values.
• Sehemu kubwa ya model parameters ni specific kwa China.
• Source code na YALMIP model files hazijashirikiwa.

Matokeo yanayoungwa mkono na utafiti

• Layered re-optimization ilipunguza day-ahead forecast cost katika real-time kwenye example iliyochaguliwa.
• Intraday na real-time electricity markets zilicheza critical role katika kusahihisha forecast deviations.
• Kutegemea day-ahead schedule pekee kulisababisha renewable-energy curtailment na load loss.
• Carbon market ilitoa important revenue chini ya free-quota structure iliyochaguliwa.
• Ancillary-service revenue ilionyesha limited lakini positive effect kwenye total cost.
• Battery ilijibu price na balance changes katika finer time scale.
• Real-time adjustment ya carbon-capture consumption ilibadilisha net gas-turbine power.
• Kuhamisha load kwanza na kisha kuipunguza kulitoa multi-layer demand response.
• Base scenario ilizalisha lower cost kuliko scenarios zote zilizolinganishwa katika real-time stage.

Matokeo ambayo utafiti haujathibitisha

• Modeli haitoi guarantee ya %21,82 savings katika real energy park.
• Haiwezi kusemwa kwamba kushiriki katika carbon market kutapunguza cost kwa %37,70 katika market designs zote.
• Free carbon quota haigawiwi kwa namna ileile katika kila nchi au kila industrial park.
• Modeli haithibitishi voltage na line security ya grid.
• Haijaonyeshwa kwa uhuru kwamba P2G system ni economically au technically profitable.
• Actual operating efficiency na long-term cost ya carbon-capture system hazijathibitishwa experimentally.
• Haijaonyeshwa kwamba operating pattern ya battery katika modeli haitaharibu cycle life.
• Haijathibitishwa kwamba forecast data zilizotengenezwa kwa normal distribution zinawakilisha real wind, solar na load errors.
• Matokeo hayawezi kuhamishwa moja kwa moja kwa electricity na carbon markets za Uturuki.
• Modeli katika hali yake ya sasa si commercial energy-management system iliyo tayari kwa field deployment.

Mbinu na Matokeo ya Utafiti

Numerical workflow

1. Decision na time-scale coupling ya electricity na carbon markets ilifafanuliwa.
2. Wind, photovoltaic, gas-turbine na carbon-capture models ziliundwa.
3. Limits za electricity purchase na sale kutoka grid zilifafanuliwa.
4. Base, shiftable na reducible load models ziliundwa.
5. Battery energy balance na charge-discharge limits ziliongezwa.
6. P2G methane production na carbon material balance zilimodeliwa.
7. Day-ahead stage ilitatuliwa kwa saa 24 na resolution ya saa 1.
8. Sehemu muhimu ya day-ahead decisions ilihamishwa kama fixed input kwenye intraday model.
9. Intraday model ilitatuliwa kwa rolling window ya saa 2 na steps za dakika 15.
10. Sehemu muhimu ya intraday decisions ilihamishwa kwenye real-time model.
11. Real-time model ilitatuliwa kwa window ya dakika 15 na steps za dakika 5.
12. Four market-participation scenarios zililinganishwa.
13. Katika kila stage, electricity operation, cost components na carbon-quota flows zilihesabiwa.

Solution tools

Cost change kati ya stages

Mabadiliko ya carbon results kati ya stages

Intraday emission ni takriban %39,31 lower kuliko day-ahead value. Real-time emission ni takriban %33,03 lower kuliko day-ahead value lakini higher kuliko intraday value. Fluctuation hii inaonyesha kwamba carbon result huhesabiwa upya katika kila stage kwa electricity purchase, gas-turbine generation na load decisions.

Technical interpretation ya comparison scenarios

Kuondoa ancillary-services market: Kwa sababu gas turbine haikuweza kupata reserve-capacity revenue, operating strategy ikawa more conservative; market electricity transactions zikachukua sehemu ya flexibility value ya gas turbine. Cost effect ilibaki limited.

Kuondoa carbon market: Excess carbon-quota sale revenue ilitoweka, na economic ordering ikawa dependent zaidi kwa short-term electricity prices. Grid electricity purchase iliongezeka huku generation yenye carbon capture ikipungua.

Kuondoa intraday na real-time electricity markets: Forecast errors zilijaribiwa kutatuliwa kwa battery, load na internal generation resources pekee. Flexibility hii haikutosha; renewable curtailment na unmet load zote zikatokea.

General scientific evaluation

Mchango mkuu wa kisayansi wa utafiti ni kugawa maamuzi katika time layers badala ya kujaribu kuboresha kwa wakati mmoja energy resources zinazohama kwa speeds tofauti. Slow variables kama gas-turbine on-off decision zinasimamiwa day-ahead, reducible load na generation deviations intraday, na fast variables kama battery na carbon-capture consumption real-time.

Layered structure hii inaruhusu modeli kusawazisha forecast accuracy na forward planning. Day-ahead plan ina cover future lakini error margin inabaki high; real-time plan inatumia more accurate information lakini inaona short horizon pekee.

Economic results za utafiti zinaonyesha kwamba carbon-quota sale revenue na multi-stage electricity trading ziliunda strong value katika example system. Hata hivyo, ukubwa wa matokeo unategemea selected free-quota na price structure.

Kwa hiyo, utafiti unapaswa kutathminiwa zaidi kama mathematical example inayoonyesha jinsi markets tofauti na time scales tofauti zinaweza kuunganishwa katika optimization system moja, badala ya definite market-design recommendation.

Maelezo ya Chanzo na Mbinu

Kichwa kamili cha asili cha utafiti: A Multi-time Scale Operation Optimization Model for Source-Grid-Load-Storage Considering Coupling of Electricity and Carbon Trading

Waandishi na mpangilio wao: Xinyue Dong; Xudong Li; Qingbo Tan; Zhuning Wang; Zhongfu Tan.

Co-first author: Hakuna tamko la equal contribution au co-first authorship katika utafiti.

Corresponding author: Xudong Li. Utafiti unatoa contact address 1169771877@qq.com. Rekodi ya SSRN pia inaonyesha Xudong Li kama contact author.

Author-institution mappings:

• Xinyue Dong, Zhuning Wang na Zhongfu Tan: School of Economics and Management, North China Electric Power University, Beijing 102206, China.
• Xudong Li: Beijing Huairou Lab, Beijing 101400, China.
• Qingbo Tan: Beijing Institute of Technology, Beijing 100081, China.

Institution metadata note: Rekodi ya SSRN haitoi institution ya Xudong Li. Beijing Huairou Lab affiliation imechukuliwa kutoka author-institution line ya utafiti.

DOI: 10.2139/ssrn.6935080

Aina ya chanzo: Preprint research article yenye mathematical operation optimization na example simulation.

Jukwaa la uchapishaji: SSRN.

Tarehe ya uchapishaji: Kulingana na rekodi ya SSRN, 13 Juni 2026.

Idadi ya kurasa: 38.

Jarida: Hakuna peer-reviewed journal name au accepted journal version iliyotajwa katika utafiti.

Mchapishaji: Hakuna separate final journal publisher; utafiti umechapishwa kwenye SSRN kama early research version.

Hali ya peer review: Utafiti haujapitiwa na rika. Kila ukurasa una warning “Preprint not peer reviewed”.

Rekodi rasmi:Ukurasa rasmi wa rekodi ya SSRN

Kiungo cha DOI:10.2139/ssrn.6935080

Funding: Utafiti uliungwa mkono na project 25CJL064 ya National Social Science Fund of China yenye jina “Collaborative Carbon Reduction Mechanism of China’s Electricity Industry in the Context of Global Carbon Neutrality”.

Conflict of interest: Waandishi walitangaza kwamba hakuna commercial au relational conflict of interest.

Ethical approval: Imeelezwa kuwa not applicable.

Author contributions:

• Xinyue Dong: Conceptualization na first draft.
• Xudong Li: Method na conceptualization.
• Qingbo Tan: Data curation.
• Zhuning Wang: Review na editing.
• Zhongfu Tan: Review na editing.

Data na code availability: Scripts zilizotengeneza forecast data, YALMIP model, Gurobi configuration au open-data repository link hazijatolewa katika utafiti.

Content hii ya Kituruki imeandaliwa kwa kuchunguza maandishi ya utafiti wa kurasa 38 uliopakiwa, equations, tables, system diagrams, market flows, carbon-quota diagrams na operation graphs. Scientific method, model parameters na results zinategemea tu taarifa zilizowasilishwa katika utafiti. External sources zilitumika tu kwa bibliographic verification ya DOI, SSRN record date, contact author na institutional identity.

Baadhi ya maelezo yaliyoandikwa baada ya cost tables katika utafiti yanasema kinyume cha table values. Aidha, ratio ya %21,82 si experimental difference kati ya proposed method na independent baseline method; ni difference kati ya day-ahead na real-time calculations za scenario ileile. Inconsistencies hizi na interpretation limits hazijasahihishwa kimya kimya, bali zimeelezwa wazi katika content.