"""MILP dispatch: dvouúrovňové SoC a záporná nákupní cena (bez DB).""" from __future__ import annotations import unittest from datetime import datetime, timedelta, timezone from types import SimpleNamespace from services.planning_engine import ( PlanningSlot, _dynamic_arb_floor_wh_series, _prewindow_deferral_slots, _slots_until_buy_le_threshold, _slots_until_sell_lt, _soc_panel_min_wh_series, solve_dispatch, ) def _slot( *, load: int = 2000, buy: float = 3.0, sell: float = 3.0, pv_a: int = 0, pv_b: int = 0, ) -> PlanningSlot: return PlanningSlot( interval_start=datetime(2026, 4, 3, 12, 0, tzinfo=timezone.utc), buy_price=buy, sell_price=sell, pv_a_forecast_w=pv_a, pv_b_forecast_w=pv_b, load_baseline_w=load, ev1_connected=False, ev2_connected=False, is_predicted_price=False, ) def _battery( *, uc_wh: float = 100_000.0, min_pct: float = 10.0, arb_pct: float = 20.0, max_pct: float = 95.0, terminal_soc_value_factor: float = 0.9, ) -> SimpleNamespace: uc = uc_wh min_wh = min_pct / 100.0 * uc arb_wh = arb_pct / 100.0 * uc return SimpleNamespace( usable_capacity_wh=uc, min_soc_wh=min_wh, arb_floor_wh=arb_wh, reserve_soc_wh=arb_wh, soc_max_wh=max_pct / 100.0 * uc, charge_efficiency=0.95, discharge_efficiency=0.95, degradation_cost_czk_kwh=0.15, max_charge_power_w=10_000, max_discharge_power_w=10_000, planner_terminal_soc_value_factor=terminal_soc_value_factor, ) class SlotsUntilSellNegativeTests(unittest.TestCase): def test_slots_until_first_negative_sell(self) -> None: base = datetime(2026, 4, 3, 0, 0, tzinfo=timezone.utc) slots: list[PlanningSlot] = [] for i in range(10): slots.append( PlanningSlot( interval_start=base + timedelta(minutes=15 * i), buy_price=1.0, sell_price=2.0 if i < 4 else -0.5, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, ) ) dist = _slots_until_sell_lt(slots, 0.0) self.assertEqual(dist[0], 4) self.assertEqual(dist[3], 1) self.assertEqual(dist[4], 0) def test_prewindow_deferral_prefers_sell_anchor(self) -> None: """Když existuje záporný prodej, kotva je vzdálenost k němu, ne k extrémnímu buy.""" base = datetime(2026, 4, 3, 0, 0, tzinfo=timezone.utc) slots: list[PlanningSlot] = [] for i in range(8): slots.append( PlanningSlot( interval_start=base + timedelta(minutes=15 * i), buy_price=-50.0, sell_price=1.0 if i < 2 else -0.1, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, ) ) adv = _prewindow_deferral_slots(slots, -2.0) self.assertEqual(adv[0], 2) def test_prewindow_deferral_falls_back_to_buy_when_no_negative_sell(self) -> None: base = datetime(2026, 4, 3, 0, 0, tzinfo=timezone.utc) slots: list[PlanningSlot] = [] for i in range(10): slots.append( PlanningSlot( interval_start=base + timedelta(minutes=15 * i), buy_price=3.0 if i < 7 else -10.0, sell_price=2.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, ) ) adv = _prewindow_deferral_slots(slots, -2.0) self.assertEqual(adv[0], 7) class SlotsUntilBuyExtremeTests(unittest.TestCase): def test_slots_until_first_extreme(self) -> None: base = datetime(2026, 4, 3, 0, 0, tzinfo=timezone.utc) slots: list[PlanningSlot] = [] for i in range(10): slots.append( PlanningSlot( interval_start=base + timedelta(minutes=15 * i), buy_price=1.0, sell_price=1.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, ) ) slots[-1] = PlanningSlot( interval_start=slots[-1].interval_start, buy_price=-10.0, sell_price=0.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, ) dist = _slots_until_buy_le_threshold(slots, -2.0) self.assertEqual(dist[0], 9) self.assertEqual(dist[8], 1) self.assertEqual(dist[9], 0) def test_prewindow_clamps_relaxed_floor_until_close(self) -> None: sm = [5000.0] * 10 dist = [9, 8, 7, 6, 5, 4, 3, 2, 1, 0] # obecná kotva (sell nebo buy) panel = _soc_panel_min_wh_series(sm, dist, 10_000.0, 20_000.0, 2) self.assertEqual(panel[0], 20_000.0) self.assertEqual(panel[6], 20_000.0) self.assertEqual(panel[7], 5000.0) self.assertEqual(panel[9], 5000.0) class DynamicArbFloorTests(unittest.TestCase): def test_more_pv_ahead_lowers_floor(self) -> None: """Čím víc FVE ve lookahead, tím nižší ekonomická podlaha v prvním slotu.""" min_w = 1_000.0 base_w = 2_000.0 uc = 10_000.0 s0 = _slot() s_low_pv = replace_slot(s0, pv_a=100, pv_b=0) s_high_pv = replace_slot(s0, pv_a=50_000, pv_b=0) ser_low = _dynamic_arb_floor_wh_series([s_low_pv] * 40, min_w, base_w, uc) ser_high = _dynamic_arb_floor_wh_series([s_high_pv] * 40, min_w, base_w, uc) self.assertLess(ser_high[0], ser_low[0]) self.assertGreaterEqual(ser_low[0], min_w) self.assertLessEqual(ser_low[0], base_w) def replace_slot( s: PlanningSlot, *, pv_a: int | None = None, pv_b: int | None = None, load: int | None = None, ) -> PlanningSlot: return PlanningSlot( interval_start=s.interval_start, buy_price=s.buy_price, sell_price=s.sell_price, pv_a_forecast_w=pv_a if pv_a is not None else s.pv_a_forecast_w, pv_b_forecast_w=pv_b if pv_b is not None else s.pv_b_forecast_w, load_baseline_w=load if load is not None else s.load_baseline_w, ev1_connected=s.ev1_connected, ev2_connected=s.ev2_connected, is_predicted_price=s.is_predicted_price, ) class PlanningDispatchMilpTests(unittest.TestCase): def test_neg_sell_with_future_neg_buy_prefers_curtail_pv_a_over_export(self) -> None: """ Když: - aktuální slot má sell < 0 (export je náklad), - v horizontu existuje budoucí buy < 0, - a zároveň existuje PV B (necurtailable) někde v horizontu, solver preferuje curtail PV A (ca) místo placeného exportu ge. """ slots = [ _slot(load=0, buy=3.0, sell=-0.1, pv_a=5000, pv_b=0), _slot(load=0, buy=-10.0, sell=1.0, pv_a=0, pv_b=5000), ] battery = _battery(uc_wh=50_000.0) hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=20_000, max_export_power_w=20_000) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.50 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 2) # Slot 0: PV A se má raději uříznout než vyvážet za zápornou cenu. self.assertEqual(int(results[0].pv_a_curtailed_w), 5000) self.assertGreaterEqual(int(results[0].grid_setpoint_w), 0) def test_two_tier_soc_solves_optimal(self) -> None: slots = [_slot()] battery = _battery() hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=15_000, max_export_power_w=15_000) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.15 * battery.usable_capacity_wh results, ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertGreaterEqual(ms, 0) self.assertEqual(len(results), 1) def test_deep_discharge_allows_covering_load_only(self) -> None: slots = [ _slot(load=3000, buy=1.0, sell=6.0, pv_a=0, pv_b=0), _slot(load=3000, buy=1.0, sell=6.0, pv_a=0, pv_b=0), ] battery = _battery(uc_wh=50_000.0) hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=20_000, max_export_power_w=20_000) vehicles = [ SimpleNamespace( max_charge_power_w=11_000, battery_capacity_kwh=50.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=11_000, battery_capacity_kwh=50.0, default_target_soc_pct=80.0, ), ] soc0 = 0.12 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 2) def test_negative_buy_price_allows_import_for_baseline(self) -> None: slots = [_slot(load=6000, buy=-0.5, sell=2.0)] battery = _battery() hp = SimpleNamespace( rated_heating_power_w=8000, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=25_000, max_export_power_w=15_000) vehicles = [ SimpleNamespace( max_charge_power_w=11_000, battery_capacity_kwh=50.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=11_000, battery_capacity_kwh=50.0, default_target_soc_pct=80.0, ), ] soc0 = 0.5 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertGreaterEqual(results[0].grid_setpoint_w, 0) def test_export_implies_end_soc_at_least_reserve(self) -> None: """Bez arbitrážní relaxace: při ge >= 1 W musí koncové soc[t] >= arb_base_wh (rezerva z DB).""" slots = [ _slot(load=500, buy=2.0, sell=8.0, pv_a=0, pv_b=0), _slot(load=500, buy=2.0, sell=8.0, pv_a=0, pv_b=0), ] battery = _battery(uc_wh=100_000.0, min_pct=10.0, arb_pct=20.0) hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=50_000, max_export_power_w=50_000) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.22 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) reserve_pct = 20.0 for r in results: if r.grid_setpoint_w < 0: self.assertGreaterEqual( r.battery_soc_target, reserve_pct - 0.2, msg="export slot must end at or above reserve SoC", ) def test_export_before_extreme_negative_buy_can_end_below_reserve(self) -> None: """ Při relaxovaném soc_min (záporný buy v lookahead) smí významný export skončit u planner floor, ne u provozní rezervy — jinak nejde ráno vypustit do sítě a nachystat kapacitu před levným nákupem. """ base = datetime(2026, 4, 3, 6, 0, tzinfo=timezone.utc) s0 = PlanningSlot( interval_start=base, buy_price=2.5, sell_price=2.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=400, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=True, ) s1 = PlanningSlot( interval_start=base + timedelta(minutes=15), buy_price=-12.0, sell_price=-0.5, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=400, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=True, ) slots = [s0, s1] battery = _battery(uc_wh=10_000.0, min_pct=10.0, arb_pct=20.0) battery.planner_extreme_buy_threshold_czk_kwh = -2.0 battery.planner_discharge_floor_percent = 5.0 battery.max_charge_power_w = 50_000 battery.max_discharge_power_w = 50_000 hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=50_000, max_export_power_w=50_000) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.88 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 2) if results[0].grid_setpoint_w < 0: self.assertLess( results[0].battery_soc_target, 19.0, msg="with relaxed soc_min, first-slot export should be able to finish below reserve %", ) def test_negative_sell_forbids_battery_export_arbitrage(self) -> None: """ Pokud sell < 0, solver nesmí vybíjet baterii do sítě pro arbitráž (dump musí proběhnout předtím). V okně sell<0 smí export vzniknout jen z přebytku FVE; zde ale FVE=0, takže očekáváme grid_setpoint>=0. """ base = datetime(2026, 4, 3, 6, 0, tzinfo=timezone.utc) s0 = PlanningSlot( interval_start=base, buy_price=2.0, sell_price=2.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=True, ) s1 = PlanningSlot( interval_start=base + timedelta(minutes=15), buy_price=2.0, sell_price=-0.5, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=True, ) s2 = PlanningSlot( interval_start=base + timedelta(minutes=30), buy_price=-15.0, sell_price=-1.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=True, ) slots = [s0, s1, s2] battery = _battery(uc_wh=10_000.0, min_pct=10.0, arb_pct=20.0) battery.planner_extreme_buy_threshold_czk_kwh = -2.0 battery.planner_discharge_floor_percent = 5.0 battery.max_charge_power_w = 50_000 battery.max_discharge_power_w = 50_000 hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=50_000, max_export_power_w=50_000) vehicles = [ SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), ] soc0 = 0.9 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 3) # V sell<0 slotu bez FVE a bez zátěže nesmí být export (to by muselo být z baterie). self.assertGreaterEqual(results[1].grid_setpoint_w, 0) # A zároveň nesmí být baterie ve výboji (dump musí proběhnout předtím). self.assertGreaterEqual(results[1].battery_setpoint_w, 0) def test_anchor_hits_floor_before_first_negative_sell(self) -> None: """ Pokud se v horizontu objeví první sell<0 a současně existuje planner floor (relaxace), solver má skončit už v předchozím slotu u planner floor (cca 5 %), ne na ~15 %. """ base = datetime(2026, 4, 3, 6, 0, tzinfo=timezone.utc) # Slot 0-1: sell >= 0; slot 2: první sell < 0; slot 3: extrémně záporný buy (motivace k bufferu). slots = [ PlanningSlot( interval_start=base, buy_price=3.0, sell_price=1.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), PlanningSlot( interval_start=base + timedelta(minutes=15), buy_price=3.0, sell_price=0.5, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), PlanningSlot( interval_start=base + timedelta(minutes=30), buy_price=3.0, sell_price=-0.2, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), PlanningSlot( interval_start=base + timedelta(minutes=45), buy_price=-20.0, sell_price=-1.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), ] battery = _battery(uc_wh=20_000.0, min_pct=10.0, arb_pct=20.0) battery.planner_extreme_buy_threshold_czk_kwh = -2.0 battery.planner_discharge_floor_percent = 5.0 battery.max_charge_power_w = 50_000 battery.max_discharge_power_w = 50_000 hp = SimpleNamespace(rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0) grid = SimpleNamespace(max_import_power_w=50_000, max_export_power_w=50_000) vehicles = [ SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), ] soc0 = 0.9 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) # Slot index 1 je poslední před prvním sell<0 (index 2). self.assertLessEqual( results[1].battery_soc_target, 6.0, msg="anchor should drive SoC close to planner floor before first negative sell", ) def test_anchor_uses_planner_floor_even_without_extreme_buy(self) -> None: """ Regrese: pokud v horizontu není buy <= threshold (soc_min_series by se nerelaxovala), kotva před sell<0 má stejně mířit na planner floor (5 %), ne na base min SoC. """ base = datetime(2026, 4, 3, 6, 0, tzinfo=timezone.utc) slots = [ PlanningSlot( interval_start=base, buy_price=3.0, sell_price=1.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), PlanningSlot( interval_start=base + timedelta(minutes=15), buy_price=3.0, sell_price=0.5, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), PlanningSlot( interval_start=base + timedelta(minutes=30), buy_price=3.0, sell_price=-0.2, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, allow_charge=True, allow_discharge_export=True, ), ] battery = _battery(uc_wh=20_000.0, min_pct=12.0, arb_pct=20.0) battery.planner_extreme_buy_threshold_czk_kwh = -2.0 battery.planner_discharge_floor_percent = 5.0 battery.max_charge_power_w = 50_000 battery.max_discharge_power_w = 50_000 hp = SimpleNamespace(rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0) grid = SimpleNamespace(max_import_power_w=50_000, max_export_power_w=50_000) vehicles = [ SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), ] soc0 = 0.9 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) # Slot index 1 je poslední před prvním sell<0 (index 2). self.assertLessEqual(results[1].battery_soc_target, 6.0) def test_grid_import_soft_cap_penalizes_breaker_overdraw(self) -> None: """ Soft cap: solver může nominálně překročit breaker, ale jen pokud se to vyplatí. Při běžné (nezáporné) nákupní ceně by měl držet import <= breaker. """ slots = [_slot(load=3700, buy=0.4, sell=-0.3, pv_a=0, pv_b=1500)] battery = _battery(uc_wh=64_000.0, min_pct=12.0, arb_pct=20.0) battery.max_charge_power_w = 18_000 battery.max_discharge_power_w = 18_000 hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=17_000, max_export_power_w=13_500) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.55 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 1) self.assertLessEqual( results[0].grid_setpoint_w, grid.max_import_power_w, msg="soft cap: for normal buy price, planned grid import should not exceed breaker", ) def test_grid_import_soft_cap_allows_overdraw_when_extremely_negative(self) -> None: """ Regrese: při extrémně záporné nákupní ceně může solver překročit breaker (za cenu penalizace), aby stihl krátké okno nabíjení. Překročení nesmí být 'zadarmo' (kontrolujeme alespoň, že existuje). """ # Dvouslotový scénář: v 1. slotu extrémně záporná cena, ve 2. slotu drahá. # Terminal SoC kotva pak nepenalizuje držení energie (průměrná buy je ~0) a solver má motivaci # v 1. slotu nabít na max, i kdyby to znamenalo malé překročení breakeru. s0 = _slot(load=0, buy=-20.0, sell=-0.3, pv_a=0, pv_b=0) s1 = replace_slot(s0, load=0) s1 = PlanningSlot( interval_start=s0.interval_start + timedelta(minutes=15), buy_price=20.0, sell_price=-0.3, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=0, ev1_connected=False, ev2_connected=False, is_predicted_price=False, ) slots = [s0, s1] battery = _battery(uc_wh=64_000.0, min_pct=12.0, arb_pct=20.0) battery.max_charge_power_w = 18_000 battery.max_discharge_power_w = 18_000 hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=17_000, max_export_power_w=13_500) vehicles = [ SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), SimpleNamespace(max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0), ] soc0 = 0.55 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 2) self.assertGreater( results[0].grid_setpoint_w, grid.max_import_power_w, msg="with very negative buy price, solver may choose to exceed breaker (soft cap)", ) def test_block_export_on_negative_sell_no_grid_export_pv_surplus(self) -> None: """site_grid_connection.block_export_on_negative_sell → ge=0 při sell<0.""" slots = [ PlanningSlot( interval_start=datetime(2026, 4, 3, 12, 0, tzinfo=timezone.utc), buy_price=5.25, sell_price=-0.5, pv_a_forecast_w=7000, pv_b_forecast_w=0, load_baseline_w=500, ev1_connected=False, ev2_connected=False, is_predicted_price=False, allow_charge=True, allow_discharge_export=False, ) ] battery = _battery(uc_wh=20_000.0, arb_pct=15.0, max_pct=95.0) hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace( max_import_power_w=17_000, max_export_power_w=8000, block_export_on_negative_sell=True, ) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.34 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 1) self.assertGreaterEqual(results[0].grid_setpoint_w, 0, "no grid export") self.assertGreater(results[0].battery_setpoint_w, 0, "surplus PV should charge") class TerminalSocShadowTests(unittest.TestCase): """Terminal SoC shadow price v objective drží konec horizontu nad holým minimem.""" def test_terminal_soc_shadow_price_prevents_drain(self) -> None: base = datetime(2026, 4, 3, 12, 0, tzinfo=timezone.utc) slots = [] for i in range(3): slots.append( PlanningSlot( interval_start=base + timedelta(minutes=15 * i), buy_price=2.0, sell_price=0.6, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=600, ev1_connected=False, ev2_connected=False, is_predicted_price=False, ) ) slots.append( PlanningSlot( interval_start=base + timedelta(minutes=45), buy_price=2.0, sell_price=14.0, pv_a_forecast_w=0, pv_b_forecast_w=0, load_baseline_w=600, ev1_connected=False, ev2_connected=False, is_predicted_price=False, ) ) battery = _battery(uc_wh=12_000.0, min_pct=12.0, arb_pct=20.0) hp = SimpleNamespace( rated_heating_power_w=0, tuv_min_temp_c=45.0, tuv_target_temp_c=55.0, ) grid = SimpleNamespace(max_import_power_w=20_000, max_export_power_w=20_000) vehicles = [ SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), SimpleNamespace( max_charge_power_w=0, battery_capacity_kwh=1.0, default_target_soc_pct=80.0, ), ] soc0 = 0.5 * battery.usable_capacity_wh results, _ms = solve_dispatch( slots, battery, hp, grid, [None, None], vehicles, soc0, 50.0, tuv_delta_stats=None, operating_mode="AUTO", ) self.assertEqual(len(results), 4) # Bez shadow price by solver mohl končit u min SoC; kotva drží znatelnou rezervu. self.assertGreaterEqual( results[-1].battery_soc_target, 15.0, msg="terminal SoC shadow price should keep end-of-horizon SoC above bare minimum", ) if __name__ == "__main__": unittest.main()