Input dataset and reproducible generator for: Carbon and Economic Returns of Risk-Constrained Stochastic Predictive Dispatch in Grid-Connected Photovoltaic, Wind, and Battery Systems

This deposit contains the complete input dataset and a reproducible generator for a study of risk-constrained stochastic predictive dispatch in a grid-connected hybrid renewable plant comprising a 50 kW photovoltaic array, a 60 kW wind turbine and a 100 kWh lithium-ion battery operating under a four-period time-of-use tariff. The generator, HRES_Dataset_Generator.m, regenerates every exogenous input to the study: the ground-truth photovoltaic, wind and demand profiles for three weather regimes; the tariff with its period index; the export price; the diurnal grid carbon intensity; and the Monte Carlo forecast uncertainty ensembles. The profiles reproduce bit-for-bit under the seeding protocol documented in the README. It requires base MATLAB R2024a with no toolboxes. The deposit also contains the numerical results underlying every table and figure of the associated article: ten comma-separated result files covering comparative performance, paired significance tests, annualised techno-economics, scenario reduction fidelity, forecast-error stress response, non-anticipativity ablation, carbon accounting under two conventions, the cost-carbon frontier, risk-measure degeneracy and the lifecycle carbon assessment; the per-seed campaign record; and thirty-four vector figures. All profiles are synthetic and are constructed to span surplus-dominated, deficit-dominated and volatility-dominated operation rather than to reproduce any specific location. The export price and the off-peak import tariff are both 0.08 EUR/kWh, a coincidence that is the subject of the associated study. # Input dataset and reproducible generator "Title : Carbon and Economic Returns of Risk-Constrained StochasticPredictive Dispatch in Grid-Connected Photovoltaic, Wind, andBattery Systems" DOI: https://doi.org/10.5281/zenodo.21874939Licence: CC BY 4.0MATLAB R2024a, base MATLAB, no toolboxes required. ## Scope This deposit contains every exogenous input consumed by the associated study,together with the numerical results underlying each of its tables and figures.The dispatch formulation itself is specified in full in Section 2 of thearticle, in numbered equations together with Algorithm 1, the decision-vectorlayout, and the constraint-matrix dimensions and sparsity, and is notreproduced here. ## Reproducing the inputs Run `HRES_Dataset_Generator.m`. It regenerates the ground-truth generation anddemand profiles for the three weather regimes, the four-period tariff with itsperiod index, the export price, the diurnal grid carbon intensity, the MonteCarlo forecast ensembles, and the parameter table. The random stream is seeded once with `rng(2026,'twister')` and the threeregimes are generated in the order clear, cloudy, windy. That order isload-bearing: generating them in a different order, or reseeding between them,yields different profiles. Forecast ensembles reseed per campaign with`rng(seed,'twister')`, so an ensemble is identified by the pair (seed, epoch).The study uses seeds 1 to 30 for each regime. ## Contents | File | Description ||---|---|| `HRES_Dataset_Generator.m` | dataset generator, 442 lines || `parameters.csv` | every parameter with value, unit and description || `signals_tariff_carbon.csv` | tariff, period index, export price, carbon intensity || `profiles_<regime>.csv` | ground-truth PV, wind and load, 192 rows at 0.25 h || `ensemble_<regime>_seed01_epoch<k>.csv` | 500 forecast trajectories, 72 columns || `Table1_Comparative.csv` | comparative performance by regime || `Table2_Significance.csv` | paired signed-rank tests with Holm correction || `Table3_TechnoEconomic.csv` | annualised techno-economic assessment || `Table4_Wasserstein.csv` | reduction fidelity by method and cardinality || `Table5_Stress.csv` | forecast-error stress response || `Table6_NA_Ablation.csv` | non-anticipativity scope ablation || `Table7_CarbonConventions.csv` | carbon account under both conventions || `Table8_CarbonFrontier.csv` | cost-carbon frontier || `Table9_RiskDegeneracy.csv` | degeneracy of the risk measure || `Table10_Lifecycle.csv` | lifecycle-adjusted carbon assessment || `SMPC_Q1_Results.mat` | per-seed campaign record, all controllers and regimes || `figures.zip` | 34 vector figures corresponding to the article | Column dictionaries for the generated files are in `README_dataset.txt`,written by the generator. ## Note on the tariff The export price and the off-peak import tariff are both 0.08 EUR/kWh. Thatcoincidence is the degeneracy the article analyses, so it is a property of thetariff rather than an oversight. ## Limitation All profiles are synthetic. They span three qualitatively distinct dispatchproblems, namely surplus-dominated, deficit-dominated and volatility-dominatedoperation, rather than reproducing any specific location. The carbon intensityis an average-factor model. ## Citation Tegani, I.; Afghoul, H.; Alharbi, S.S.; Alharbi, S.; Tegani, S. Input datasetand reproducible generator for a risk-constrained stochastic predictivedispatch study. Zenodo, 2026. https://doi.org/10.5281/zenodo.21874939