def process_request(self, fleet_request):
"""
The expectation that configuration will have at least the following
items
:param fleet_request: an instance of FleetRequest
:return res: an instance of FleetResponse
"""
res = FleetResponse()
# Dummy values for testing
res.P_togrid = 100
res.P_service = 100
res.Q_service = 100
res.Q_togrid = 100
return res
def run(self, P_req=[0], Q_req=[0],ts=datetime.utcnow(), del_t=timedelta(hours=1)):
'''
This function takes the real and reactive power requested, date, time, and time step and
calculates the updated fleet variables as well as a FleetResponse object based on what the
simulated fleet is able to provide. It does this by dividing up the requests and sending each
device in the fleet its own request. If some of the fleet is unable to supply the requested
power, then the remainder is devided umung the remaining devices.
:param P_req: requested real power, Q_req: requested reactive power
ts:datetime opject, del_t: timedelta object
:return fleet_response: an instance of FleetResponse
'''
np = numpy.ones(self.num_of_devices,int)
nq = numpy.ones(self.num_of_devices,int)
p_none = 0
q_none = 0
if P_req==None:
P_req = 0
p_req = 0
p_none = 1
else:
p_req = P_req/self.num_of_devices
if Q_req==None:
Q_req = 0
q_req = 0
q_none = 1
else:
q_req = Q_req/self.num_of_devices
p_tot = 0.0
q_tot = 0.0
dt = del_t.total_seconds() / 3600.0
self.t = self.t + dt
p_FW = numpy.zeros(self.num_of_devices,float)
q_VV = numpy.zeros(self.num_of_devices,float)
response = FleetResponse()
response.ts = ts
last_P = numpy.zeros(self.num_of_devices,int)
last_Q = numpy.zeros(self.num_of_devices,int)
soc_update = copy.copy(self.soc)
if self.model_type == 'CRM':
pdc_update = self.pdc
ibat_update = self.ibat
v1_update = self.v1
v2_update = self.v2
vbat_update = self.vbat
for i in range(self.num_of_devices):
last_P[i] = self.P_service[i]
last_Q[i] = self.Q_service[i]
self.P_service[i] = 0
self.Q_service[i] = 0
TOL = 0.000001 # tolerance
one_pass = 1 #guarantees at least one pass through the control loop
while ((p_tot < P_req-TOL or p_tot > P_req+TOL) or (q_tot < Q_req-TOL or q_tot > Q_req+TOL)) and sum(np)!=0 and sum(nq)!=0 or one_pass == 1: #continue looping through devices until the power needs are met or all devices are at their limits
one_pass = 0
# distribute the requested power equally among the devices that are not at their limits
p_req = (P_req - p_tot)/sum(np)
q_req = (Q_req - q_tot)/sum(nq)
for i in range(self.num_of_devices):
if self.model_type == 'ERM':
soc_update[i] = self.run_soc_update(p_req,q_req,np,nq,last_P,last_Q,i,dt)
if self.model_type == 'CRM':
[soc_update[i],pdc_update[i],ibat_update[i],vbat_update[i],v1_update[i],v2_update[i]] = \
self.run_soc_update(p_req,q_req,np,nq,last_P,last_Q,i,dt)
# at the end of looping through the devices, add up their power to determine if the request has been met
p_tot = sum(self.P_service)
q_tot = sum(self.Q_service)
# after all the power needs have been placed and met, then make adjustments based on autonomous operation settings
if (self.FW21_Enabled == True or self.VV11_Enabled == True) and self.is_autonomous == True:
for i in range(self.num_of_devices) :
p_req = self.P_service[i]
q_req = self.Q_service[i]
# determin if VV11 or FW21 change the p_req or q_req
if self.FW21_Enabled == True:
p_mod = self.frequency_watt(p_req,ts,self.location[i])
if self.VV11_Enabled == True:
if q_none == 1:
q_mod = self.volt_var(ts,self.location[i])
if self.model_type == 'ERM':
soc_update[i] = self.run_soc_update(p_req=(p_mod-p_req),q_req=(q_mod-q_req),np=numpy.ones(self.num_of_devices),nq=numpy.ones(self.num_of_devices),last_P=last_P,last_Q=last_Q,i=i,dt=dt)
if self.model_type == 'CRM':
[soc_update[i],pdc_update[i],ibat_update[i],vbat_update[i],v1_update[i],v2_update[i]] = \
self.run_soc_update(p_req=(p_mod-p_req),q_req=(q_mod-q_req),np=numpy.ones(self.num_of_devices),nq=numpy.ones(self.num_of_devices),last_P=last_P,last_Q=last_Q,i=i,dt=dt)
else :
if self.model_type == 'ERM':
soc_update[i] = self.run_soc_update(p_req=(p_req),q_req=(q_req),np=numpy.ones(self.num_of_devices),nq=numpy.ones(self.num_of_devices),last_P=last_P,last_Q=last_Q,i=i,dt=dt)
if self.model_type == 'CRM':
[soc_update[i],pdc_update[i],ibat_update[i],vbat_update[i],v1_update[i],v2_update[i]] = \
self.run_soc_update(p_req=p_req,q_req=q_req,np=numpy.ones(self.num_of_devices),nq=numpy.ones(self.num_of_devices),last_P=last_P,last_Q=last_Q,i=i,dt=dt)
p_tot = sum(self.P_service)
q_tot = sum(self.Q_service)
# update SoC
self.soc = soc_update
# update SoH
if self.model_type == 'ERM':
self.soh = self.soh - 100*dt*abs(self.P_service)/((1+1/self.energy_efficiency)*self.cycle_life*self.energy_capacity)
if self.model_type == 'CRM':
self.v1 = v1_update
self.v2 = v2_update
self.voc_update()
self.ibat = ibat_update
self.vbat = (self.v1 + self.v2 + self.voc + self.ibat*self.r0) *self.n_cells
self.soh = self.soh - 100*dt*abs(self.ibat)/((1+1/self.coulombic_efficiency)*self.cycle_life*self.charge_capacity)
# once the power request has been met, or all devices are at their limits, return the response variables
response.P_service = p_tot
response.Q_service = q_tot
response.soc = numpy.average(self.soc)
response.E = numpy.average(self.soc) * self.energy_capacity / 100.0
return response
def run(self, P_req=[0], Q_req=[0], initSoC=50, dt=1):
P_req = P_req / self.num_of_devices
Q_req = Q_req / self.num_of_devices
# error checking
if initSoC < self.min_soc or initSoC > self.max_soc:
print('ERROR: initSoC out of range')
return [[], []]
elif P_req == []:
print('ERROR: P_req vector must not be empty')
return [[], []]
else:
self.t = self.t + dt
response = FleetResponse()
# pre-define output vectors SoC
SoC = numpy.zeros(2)
SoC[0] = initSoC # initialize SoC
p_ach = 0
q_ach = 0
# Max ramp rate and aparent powerlimit checking
if (P_req - self.P_injected) > self.max_ramp_up:
p_ach = self.max_ramp_up
elif (P_req - self.P_injected) < self.max_ramp_down:
p_ach = self.max_ramp_down
else:
p_ach = P_req
if (Q_req - self.Q_injected) > self.max_ramp_up:
q_ach = self.max_ramp_up
elif (Q_req - self.Q_injected) < self.max_ramp_down:
q_ach = self.max_ramp_down
else:
q_ach = Q_req
if p_ach < self.max_power_discharge:
p_ach = self.max_power_discharge
if p_ach > self.max_power_charge:
p_ach = self.max_power_charge
S_req = float(numpy.sqrt(p_ach**2 + q_ach**2))
if S_req > self.max_apparent_power:
q_ach = float(
numpy.sqrt(numpy.abs(self.max_apparent_power**2 -
p_ach**2)) * numpy.sign(q_ach))
S_req = float(numpy.sqrt(p_ach**2 + q_ach**2))
if p_ach != 0.0:
if float(numpy.abs(S_req / p_ach)) < self.min_pf:
q_ach = float(
numpy.sqrt(
numpy.abs((p_ach / self.min_pf)**2 - p_ach**2)) *
numpy.sign(q_ach))
# run function for ERM model type
if self.model_type == 'ERM':
response.P_injected_max = self.max_power_discharge
response.Q_injected_max = float(
numpy.sqrt(numpy.abs(self.max_apparent_power**2 -
p_ach**2)))
# Calculate SoC and Power Achieved
Ppos = min(self.max_power_charge, max(p_ach, 0))
Pneg = max(self.max_power_discharge, min(p_ach, 0))
SoC[1] = SoC[0] + float(100) * dt * (
Pneg + (Ppos * self.energy_efficiency) +
self.self_discharge_power) / self.energy_capacity
if SoC[1] > self.max_soc:
Ppos = (self.energy_capacity * (self.max_soc - SoC[0]) /
(float(100) * dt) -
self.self_discharge_power) / self.energy_efficiency
SoC[1] = self.max_soc
if SoC[0] == self.max_soc:
Ppos = 0
if SoC[1] < self.min_soc:
Pneg = self.energy_capacity * (self.min_soc - SoC[0]) / (
float(100) * dt) - self.self_discharge_power
SoC[1] = self.min_soc
if SoC[0] == self.min_soc:
Pneg = 0
p_ach = (Ppos + Pneg) * self.num_of_devices
q_ach = q_ach * self.num_of_devices
response.P_injected = p_ach
response.Q_injected = q_ach
response.P_dot = p_ach - self.P_injected
response.Q_dot = q_ach - self.Q_injected
response.P_service = 0
response.Q_service = 0
response.P_service_max = 0
response.Q_service_max = 0
response.Loss_standby = self.self_discharge_power
response.Eff_throughput = float(
p_ach > 0) * self.energy_efficiency + float(p_ach > 0)
self.soc = SoC[1]
return response
# run function for ERM model type
elif self.model_type == 'CRM':
# convert AC power p_ach to DC power pdc
self.pdc = self.coeff_2 * (p_ach**2) + self.coeff_1 * (
p_ach) + self.coeff_0
# convert DC power pdc to DC current
""" self.ibat = self.pdc *1000 / self.vbat
self.pdc = self.ibat * ((self.v1 + self.v2 + self.voc + self.ibat*self.r0) *self.n_cells) *1000
0 = -self.pdc + (self.v1 + self.v2 + self.voc)*self.n_cells) *1000*self.ibat + self.r0 *self.n_cells*1000* (self.ibat**2) """
b = ((self.v1 + self.v2 + self.voc) * self.n_cells)
a = self.r0 * self.n_cells
c = -self.pdc * 1000
self.ibat = (-b + numpy.sqrt(b**2 - 4 * a * c)) / (2 * a)
self.vbat = (self.v1 + self.v2 + self.voc +
self.ibat * self.r0) * self.n_cells
# calculate dynamic voltages
self.v1 = self.v1 + dt * ((1 / (self.r1 * self.c1)) * self.v1 +
(1 / (self.c1)) * self.ibat)
self.v2 = self.v2 + dt * ((1 / (self.r2 * self.c2)) * self.v1 +
(1 / (self.c2)) * self.ibat)
response.P_injected_max = self.max_power_discharge
response.Q_injected_max = float(
numpy.sqrt(numpy.abs(self.max_apparent_power**2 -
p_ach**2)))
# Calculate SoC and Power Achieved
Ipos = min(self.max_current_charge, max(self.ibat, 0))
Ineg = max(self.max_current_discharge, min(self.ibat, 0))
SoC[1] = SoC[0] + float(100) * dt * (
Ineg + (Ipos * self.coulombic_efficiency) +
self.self_discharge_current) / self.charge_capacity
if SoC[1] > self.max_soc:
Ipos = (self.charge_capacity * (self.max_soc - SoC[0]) /
(float(100) * dt) - self.self_discharge_current
) / self.coulombic_efficiency
SoC[1] = self.max_soc
if SoC[0] == self.max_soc:
Ipos = 0
self.pdc = Ipos * self.vbat / 1000
if self.coeff_2 != 0:
p_ach = (-self.coeff_1 + float(
numpy.sqrt(self.coeff_1**2 - 4 * self.coeff_2 *
(self.coeff_0 - self.pdc)))) / (
2 * self.coeff_2)
else:
p_ach = (self.pdc - self.coeff_0) / self.coeff_1
if SoC[1] < self.min_soc:
Ineg = self.charge_capacity * (self.min_soc - SoC[0]) / (
float(100) * dt) - self.self_discharge_current
SoC[1] = self.min_soc
if SoC[0] == self.min_soc:
Ineg = 0
self.pdc = Ineg * self.vbat / 1000
if self.coeff_2 != 0:
p_ach = (-self.coeff_1 + float(
numpy.sqrt(self.coeff_1**2 - 4 * self.coeff_2 *
(self.coeff_0 - self.pdc)))) / (
2 * self.coeff_2)
else:
p_ach = (self.pdc - self.coeff_0) / self.coeff_1
self.ibat = Ipos + Ineg
self.soc = SoC[1]
self.voc_update()
self.vbat = (self.v1 + self.v2 + self.voc +
self.ibat * self.r0) * self.n_cells
p_ach = p_ach * self.num_of_devices
q_ach = q_ach * self.num_of_devices
response.P_injected = p_ach
response.Q_injected = q_ach
response.P_dot = p_ach - self.P_injected
response.Q_dot = q_ach - self.Q_injected
response.P_service = 0
response.Q_service = 0
response.P_service_max = 0
response.Q_service_max = 0
response.Loss_standby = self.self_discharge_current * (
self.voc)
response.E = (self.soc -
self.min_soc) * self.charge_capacity * self.voc
response.Eff_throughput = (response.E - self.es) / (p_ach * dt)
self.es = response.E
self.vbat = (self.v1 + self.v2 + self.voc +
self.ibat * self.r0) * self.n_cells
return response
def run(self, P_req, Q_req, initSOC, t, dt, ts):
# ExecuteFleet(self, Steps, Timestep, P_request, Q_request, forecast):
# run(self, P_req=[0], Q_req=[0], ts=datetime.utcnow(), del_t=timedelta(hours=1)):
# Give the code the capability to respond to None requests
if P_req == None:
P_req = 0
if Q_req == None:
Q_req = 0
P_togrid = 0
P_service = 0
P_base = 0
P_service_max = 0
self.P_request_perWH = P_req / self.numWH # this is only for the first step
number = 0 # index for running through all the water heaters
NumDevicesToCall = 0
P_service_max0 = 0
# decision making about which water heater to call on for service, check if available at last step, if so then
# check for SoC > self.minSOC and Soc < self.maxSOC
for n in range(self.numWH):
if self.P_request_perWH < 0 and self.IsAvailableAdd[
n] > 0 and self.SOC[n] < self.maxSOC:
NumDevicesToCall += 1
elif self.P_request_perWH > 0 and self.IsAvailableShed[
n] > 0 and self.SOC[n] > self.minSOC:
NumDevicesToCall += 1
if P_req != None:
self.P_request_perWH = P_req / max(
NumDevicesToCall, 1
) # divide the fleet request by the number of devices that can be called upon
# create a .csv outputfile with each water heater's metrics
# outputfilename = join(self.base_path,"WH_fleet_outputs.csv")
# self.outputfile = open(outputfilename,"w")
# self.outputfile.write("Timestep,")
# create a .csv outputfile with P_service, P_togrid, and P_base
# outputfilename = join(self.base_path,"WH_fleet_outputs.csv")
#################################
for wh in self.whs: # loop through all water heaters
if P_req == None:
response = wh.execute(
self.TtankInitial[number], self.TtankInitial_b[number],
self.TsetInitial[number], self.Tamb[number][0],
self.RHamb[number][0], self.Tmains[number][0],
self.draw[number][0], 0, self.Type, self.dt,
self.draw_fleet_ave[0], self.element_on_last)
P_service = 0
if P_req < 0 and self.IsAvailableAdd[number] > 0:
response = wh.execute(
self.TtankInitial[number], self.TtankInitial_b[number],
self.TsetInitial[number], self.Tamb[number][0],
self.RHamb[number][0], self.Tmains[number][0],
self.draw[number][0], P_req, self.Type, self.dt,
self.draw_fleet_ave[0], self.element_on_last)
P_req = P_req - response.Eservice
P_service += response.Eservice
elif P_req > 0 and self.IsAvailableShed[number] > 0:
response = wh.execute(
self.TtankInitial[number], self.TtankInitial_b[number],
self.TsetInitial[number], self.Tamb[number][0],
self.RHamb[number][0], self.Tmains[number][0],
self.draw[number][0], P_req, self.Type, self.dt,
self.draw_fleet_ave[0], self.element_on_last)
P_req = P_req + response.Eservice
P_service -= response.Eservice
#print("P_req = {}, P_service = {}, Eservice = {}".format(P_req,P_service,response.Eservice))
else:
response = wh.execute(
self.TtankInitial[number], self.TtankInitial_b[number],
self.TsetInitial[number], self.Tamb[number][0],
self.RHamb[number][0], self.Tmains[number][0],
self.draw[number][0], 0, self.Type, self.dt,
self.draw_fleet_ave[0], self.element_on_last)
# print('P_req = {}'.format(P_req))
# assign returned parameters to associated lists to be recorded
self.element_on_last[number] = response.ElementOn
self.TtankInitial[number] = response.Ttank
self.TtankInitial_b[number] = response.Ttank_b
self.SOC[number] = response.SOC
self.SOCb[number] = response.SOC_b
self.IsAvailableAdd[number] = response.IsAvailableAdd
self.IsAvailableShed[number] = response.IsAvailableShed
self.AvailableCapacityAdd[number] = response.AvailableCapacityAdd
self.AvailableCapacityShed[number] = response.AvailableCapacityShed
self.ServiceCallsAccepted[number] = response.ServiceCallsAccepted
self.ServiceProvided[number] = response.Eservice
'''
P_togrid -= response.Eused
P_base -= response.Pbase
if P_req <0 or P_req > 0:
P_response = (P_togrid - P_base)
else:
P_response = 0
'''
P_togrid -= response.Eused
P_base -= response.Pbase
# self.outputfile.write(str(response.Ttank) +"," + str(self.TsetInitial[number]) + "," + str(response.Eused) + "," + str(response.PusedMax) + "," + str(response.Eloss) + "," + str(response.ElementOn) + "," + str(response.Eservice) + "," + str(response.SOC) + "," + str(response.AvailableCapacityAdd) + "," + str(response.AvailableCapacityShed) + "," + str(response.ServiceCallsAccepted) + "," + str(response.IsAvailableAdd) + "," + str(response.IsAvailableShed) + "," + str(self.draw[number][0]) + "," + str(response.Edel) + ",")
# resp.sim_step = response.sim_step
number += 1 # go to next device
if P_req <= 0:
P_service_max += response.AvailableCapacityShed # NOTE THIS ASSUMES THE MAX SERVICE IS LOAD SHED
else:
P_service_max0 += response.AvailableCapacityAdd
P_service_max = -1.0 * P_service_max0
# self.outputfile.write("\n")
self.step += 1 # To advance the step by step in the disturbance file
# Output Fleet Response
resp = FleetResponse()
resp.P_service = []
resp.P_service_max = []
resp.P_service_min = []
resp.P_togrid = []
resp.P_togrid_max = []
resp.P_togrid_min = []
resp.P_forecast = []
resp.P_base = []
resp.E = []
resp.C = []
resp.ts = ts
resp.sim_step = dt
# resp.P_dot_up = resp.P_togrid_max / ServiceRequest.Timestep.seconds
resp.P_service_max = P_service_max
resp.P_service = P_service
resp.P_base = P_base
resp.P_togrid = P_togrid
# if P_service != 0:
# print("resp.P_base = {}".format(resp.P_base))
# print("Pbase = {}".format(resp.P_base))
# print("Ptogrid = {}".format(resp.P_togrid))
# Available Energy stored at the end of the most recent timestep
# resp.E += response.Estored
resp.E = 0
resp.C += response.SOC / (self.numWH)
resp.Q_togrid = 'NA'
resp.Q_service = 'NA'
resp.Q_service_max = 'NA'
resp.Q_service_min = 'NA'
resp.Q_togrid_min = 'NA'
resp.Q_togrid_max = 'NA'
resp.Q_dot_up = 'NA'
resp.Q_dot_down = 'NA'
resp.P_dot_up = 0
resp.P_dot_down = 0
resp.Eff_charge = 1.0 # TODO: change this if we ever use a HPWH to use the HP efficiency
resp.Eff_discharge = 1.0 # always equal to 1 for this device
resp.P_dot = resp.P_togrid / dt
resp.P_service_min = 0
resp.dT_hold_limit = 'NA'
resp.T_restore = 'NA'
resp.Strike_price = 'NA'
resp.SOC_cost = 'NA'
# TotalServiceProvidedPerTimeStep[step] = -1.0*self.P_service # per time step for all hvacs -1.0*
"""
Modify the power and SOC of the different subfeets according
to the frequency droop regulation according to IEEE standard
"""
if self.FW21_Enabled and self.is_autonomous:
power_ac = resp.P_service_max
p_prev = resp.P_togrid
power_fleet = self.frequency_watt(power_ac, p_prev, self.ts,
self.location,
self.db_UF_subfleet,
self.db_OF_subfleet)
power_fleet, SOC_step = self.update_soc_due_to_frequency_droop(
initSOC, power_fleet, dt)
self.time = t + dt
self.ts = self.ts + timedelta(seconds=dt)
# Restart time if it surpasses 24 hours
if self.time > 24 * 3600:
self.time = self.time - 24 * 3600
# Impact Metrics
# Update the metrics
'''
Ttotal = 0
SOCTotal = 0
self.cycle_basee = np.sum(self.cycle_off_base)
self.cycle_basee += np.sum(self.cycle_on_base)
self.cycle_grid = np.sum(self.cycle_off_grid)
self.cycle_grid += np.sum(self.cycle_on_grid)
'''
for number in range(self.numWH):
if response.Ttank <= self.TsetInitial[
number] - 10: # assume 10F deadband (consistent with wh.py)
self.unmet_hours += 1 * self.sim_step / 3600.0
if resp.P_base == 0 and resp.P_togrid == 0:
self.ratio_P_togrid_P_base = 1.0
elif resp.P_base == 0 and resp.P_togrid != 0:
self.ratio_P_togrid_P_base = 'NA'
else:
self.ratio_P_togrid_P_base = resp.P_togrid / (resp.P_base)
self.energy_impacts += abs(resp.P_service) * (self.sim_step / 3600)
return resp
def Run_Fleet(self, ts, sim_step, P_req, Q_req, return_forecast, WP):
# this module aggregates the PV information from a large pool of PV/invertre system
# Updates prepares report for system operator based on PV fleet
# Breaks down the operation command from system operator to subfleet and device level
if P_req == None:
P_req = []
if Q_req == None:
Q_req = []
def Grid_Param():
from grid_info import GridInfo
a = GridInfo('Grid_Info_DATA_2.csv')
V = a.get_voltage('XX')
f = a.get_frequency('XX')
return f, V
# pull information about the PV subfleet
P_rated = self.p_rated
P_min = self.p_min
S_max = self.s_max
Qmax_Plus = self.qmax_plus
Qmax_minus = self.qmax_minus
P_up = self.p_ramp_down
P_down = self.p_ramp_down
Q_up = self.q_ramp_down
Q_down = self.q_ramp_down
# Create Battery equivalent model (BEM) variables for the subfleets and PV fleet
# SubFleetA_NP_C='NA'
# SubFleetA_NP_P_grid_max=P_rated*self.SubFleet_NumberOfUnits #Minimum real power for services
# SubFleetA_NP_Q_grid_max=Qmax_Plus*self.SubFleet_NumberOfUnits#Maximum reactive power for services
# SubFleetA_NP_P_grid_min=P_min*self.SubFleet_NumberOfUnits#Minimum reactive power for services
# SubFleetA_NP_Q_grid_min=Qmax_minus*self.SubFleet_NumberOfUnits#Ramp rate real power up
# SubFleetA_NP_P_service_max=SubFleetA_NP_P_grid_max-SubFleetA_NP_P_grid_min#Ramp rate real power down
# SubFleetA_NP_Q_service_max=SubFleetA_NP_Q_grid_max-SubFleetA_NP_Q_grid_min#Ramp rate reactive power up
# SubFleetA_NP_P_service_min=0#Ramp rate reactive power down
# SubFleetA_NP_Q_service_min=0
SubFleetA_NP_P_up = P_up
SubFleetA_NP_Q_up = Q_up
SubFleetA_NP_P_down = P_down
SubFleetA_NP_Q_down = Q_down
# SubFleetA_NP_e_in='NA'
SubFleetA_NP_e_out = []
SubFleetA_NP_S_rating = S_max * self.SubFleet_NumberOfUnits
Fleet_PV_NP_C = 'NA'
# Fleet_PV_NP_P_grid_max=SubFleetA_NP_P_grid_max#Minimum real power for services
# Fleet_PV_NP_Q_grid_max=SubFleetA_NP_Q_grid_max#Maximum reactive power for services
# Fleet_PV_NP_P_grid_min=SubFleetA_NP_P_grid_min#Minimum reactive power for services
# Fleet_PV_NP_Q_grid_min=SubFleetA_NP_Q_grid_min#Ramp rate real power up
# Fleet_PV_NP_P_service_max=SubFleetA_NP_P_service_max#Ramp rate real power down
# Fleet_PV_NP_Q_service_max=SubFleetA_NP_Q_service_max#Ramp rate reactive power up
# Fleet_PV_NP_P_service_min=0#Ramp rate reactive power down
# Fleet_PV_NP_Q_service_min=0
Fleet_PV_NP_P_up = SubFleetA_NP_P_up
Fleet_PV_NP_Q_up = SubFleetA_NP_Q_up
Fleet_PV_NP_P_down = SubFleetA_NP_P_down
Fleet_PV_NP_Q_down = SubFleetA_NP_Q_down
Fleet_PV_NP_e_in = 'NA'
# Fleet_PV_NP_e_out=[]
FleetA_NP_S_rating = SubFleetA_NP_S_rating
# break down the command for fleet to command for device
Command_to_Device = self.Aggregator_Command(
P_req, Q_req, self.SubFleet_NumberOfUnits)
# Grid parameters required for autonomous mode of operation
#[Time_,P_rec,Q_rec,P_req_,Q_req_,P_MPP,G_rec,T_rec]\
[P_grid,Q_grid,P_service,Q_service,E_t0,c,P_output,Q_output,P_grid_max,Q_grid_max,\
P_grid_min,Q_grid_min,P_service_max,Q_service_max,P_service_min,Q_service_min,del_t_hold,\
t_restore,SP,N_req,Effeciency,sim_step]=self.Device_PV(ts,sim_step,Command_to_Device,return_forecast,WP)
#print(P_Output)
SubFleetA_P_grid = P_grid * self.SubFleet_NumberOfUnits / 1000
SubFleetA_Q_grid = Q_grid * self.SubFleet_NumberOfUnits / 1000
SubFleetA_P_service = P_service * self.SubFleet_NumberOfUnits / 1000
SubFleetA_Q_service = Q_service * self.SubFleet_NumberOfUnits / 1000
SubFleetA_E_t0 = E_t0
SubFleetA_c = c
SubFleetA_P_output = [
x * self.SubFleet_NumberOfUnits / 1000 for x in P_output
]
SubFleetA_Q_output = [
x * self.SubFleet_NumberOfUnits / 1000 for x in Q_output
]
SubFleetA_P_grid_max = [
x * self.SubFleet_NumberOfUnits / 1000 for x in P_grid_max
]
SubFleetA_Q_grid_max = [
x * self.SubFleet_NumberOfUnits / 1000 for x in Q_grid_max
]
SubFleetA_P_grid_min = [
x * self.SubFleet_NumberOfUnits / 1000 for x in P_grid_min
]
SubFleetA_Q_grid_min = [
x * self.SubFleet_NumberOfUnits / 1000 for x in Q_grid_min
]
SubFleetA_P_service_max = [
x * self.SubFleet_NumberOfUnits / 1000 for x in P_service_max
]
SubFleetA_Q_service_max = [
x * self.SubFleet_NumberOfUnits / 1000 for x in Q_service_max
]
SubFleetA_P_service_min = [
x * self.SubFleet_NumberOfUnits / 1000 for x in P_service_min
]
SubFleetA_Q_service_min = [
x * self.SubFleet_NumberOfUnits / 1000 for x in Q_service_min
]
SubFleetA_del_t_hold = del_t_hold
SubFleetA_t_restore = t_restore
SubFleetA_SP = SP
SubFleetA_N_req = N_req
SubFleetA_NP_e_out = Effeciency
# Calculate information about Subfleets
# Calculate information about Device Fleet: PV
# Calculate information about Subfleets
Fleet_PV_P_grid = SubFleetA_P_grid
Fleet_PV_Q_grid = SubFleetA_Q_grid
Fleet_PV_P_service = SubFleetA_P_service
Fleet_PV_Q_service = SubFleetA_Q_service
Fleet_PV_E_t0 = SubFleetA_E_t0
# Fleet_PV_c=SubFleetA_c
#
# Fleet_PV_P_output=SubFleetA_P_output
# Fleet_PV_Q_output=SubFleetA_Q_output
Fleet_PV_P_grid_max = SubFleetA_P_grid_max
Fleet_PV_Q_grid_max = SubFleetA_Q_grid_max
Fleet_PV_P_grid_min = SubFleetA_P_grid_min
Fleet_PV_Q_grid_min = SubFleetA_Q_grid_min
Fleet_PV_P_service_max = SubFleetA_P_service_max
Fleet_PV_Q_service_max = SubFleetA_Q_service_max
Fleet_PV_P_service_min = SubFleetA_P_service_min
Fleet_PV_Q_service_min = SubFleetA_Q_service_min
Fleet_PV_del_t_hold = SubFleetA_del_t_hold
Fleet_PV_t_restore = SubFleetA_t_restore
# Fleet_PV_SP=SubFleetA_SP
# Fleet_PV_N_req=SubFleetA_N_req
response = FleetResponse()
response.ts = ts
response.C = Fleet_PV_NP_C
response.dT_hold_limit = Fleet_PV_del_t_hold
response.E = Fleet_PV_E_t0
response.Eff_charge = Fleet_PV_NP_e_in
response.Eff_discharge = SubFleetA_NP_e_out
response.P_dot_down = Fleet_PV_NP_P_down
response.P_dot_up = Fleet_PV_NP_P_up
response.P_service = Fleet_PV_P_service
response.P_service_max = Fleet_PV_P_service_max
response.P_service_min = Fleet_PV_P_service_min
response.P_togrid = Fleet_PV_P_grid
response.P_togrid_max = Fleet_PV_P_grid_max[0]
response.P_togrid_min = Fleet_PV_P_grid_min[0]
response.Q_dot_down = Fleet_PV_NP_Q_down
response.Q_dot_up = Fleet_PV_NP_Q_up
response.Q_service = Fleet_PV_Q_service
response.Q_service_max = Fleet_PV_Q_service_max
response.Q_service_min = Fleet_PV_Q_service_min
response.Q_togrid = Fleet_PV_Q_grid
response.Q_togrid_max = Fleet_PV_Q_grid_max[0]
response.Q_togrid_min = Fleet_PV_Q_grid_min[0]
response.S_Rating = FleetA_NP_S_rating
response.T_restore = Fleet_PV_t_restore
response.P_base = Fleet_PV_P_grid_max[0]
response.sim_step = sim_step
response.Strike_price = 0
response.SOC_cost = 'na'
self.energy_impacts += abs(
response.P_service) * (int(sim_step.total_seconds()) / 3600)
return response
def fc_model(self, ts, sim_step, Preq, forecast=False, start_time=None):
"""
:param ts: Request created for current time-step: datetime
:param sim_step: Request for simulation time-step: timedelta object
:param Preq: Request for current real power request
:param forecast: Returns fleet response forecast
:param start_time: Request for current real power request
:return resp: Fleet response object
"""
if self.P_tank_ideal <= self.min_charge:
pass
else:
# Create base power profile
if isinstance(self.p, ndarray):
self.__p_base = sum([
self.p[j] * (self.__inc + 1)**(21 - j) for j in range(22)
])
else:
self.__p_base = self.p
# Ptogrid: Pbase for None or zero requests since FC is a source
if Preq is None or float(Preq) == 0.0:
Preq = self.__p_base
# Ptogrid: Preq+Pbase for +ve or -ve requests
else:
Preq += self.__p_base
# Compute the power generated by the fleet
self.__fc_p_calc(Preq)
if self.FW21_Enabled and self.is_autonomous:
# all in kW
Preq, self.f = self.frequency_watt(p_pre=self.__p_tot_ideal,
p_avl=self.fc_Pmax_fleet,
p_min=self.fc_Pmin_fleet,
ts=ts,
start_time=start_time)
self.__p_tot_ideal = Preq
# Update SOC
self.__soc_calc()
self.__inc += 1
# Response
# Power injected is positive
resp = FleetResponse()
resp.ts = ts
resp.sim_step = sim_step
resp.C = 0
resp.dT_hold_limit = 0
resp.E = self.soc_ideal
resp.Eff_charge = 1.0
resp.Eff_discharge = self.__eta_ds
resp.P_dot_down = 0
resp.P_dot_up = 0
resp.P_togrid = self.__p_tot_ideal
resp.P_togrid_max = self.fc_Pmax_fleet
resp.P_togrid_min = self.fc_Pmin_fleet
resp.P_service = resp.P_togrid - self.__p_base
resp.P_service_max = self.fc_Pmax_fleet - self.__p_base
resp.P_service_min = self.__p_base - self.fc_Pmin_fleet
resp.Q_dot_down = 0
resp.Q_dot_up = 0
resp.Q_service = 0
resp.Q_service_max = 0
resp.Q_service_min = 0
resp.Q_togrid = 0
resp.Q_togrid_max = 0
resp.Q_togrid_min = 0
resp.T_restore = 0
resp.P_base = self.__p_base
resp.Q_base = 0
resp.Strike_price = 0
resp.SOC_cost = 0
# Impact metrics
if not forecast:
self.metrics.append([
str(ts),
str(self.__Vr),
str(self.__Vr_age),
str(self.__ne),
str(self.__ne_age),
str(self.soc_ideal * 1e2),
str(self.soc_age * 1e2),
str(self.lka_h2),
str(self.__eta_ds * 1e2)
])
# Print Soc every 5 secs.
if self.__inc % 5000 == 0:
print("Soc:%4.2f%%" % (resp.E * 1e2))
return resp
def run(self, P_req, Q_req, initSOC, t, dt, ts):
#ExecuteFleet(self, Steps, Timestep, P_request, Q_request, forecast):
# run(self, P_req=[0], Q_req=[0], ts=datetime.utcnow(), del_t=timedelta(hours=1)):
# Give the code the capability to respond to None requests
if P_req == None:
P_req = 0
if Q_req == None:
Q_req = 0
# run through fleet once as a forecast just to get initial conditions
number = 0 # run through all the devices
servsum = 0
servmax = 0
servmax2 = 0
NumDevicesToCall = 0
p_togrid = 0
p_service = 0
p_base = 0
service_max = 0
service_min = 0
# laststep = step - 1
# decision making about which RF to call on for service, check if available at last step, if so then
# check for SoC > self.minSOC and Soc < self.maxSOC
P_request_perRF = P_req/max(NumDevicesToCall,1) #divide the fleet request by the number of devices that can be called upon
for n in range(self.numRF):
if P_request_perRF < 0 and self.IsAvailableAdd[n] > 0:
NumDevicesToCall += 1
elif P_request_perRF > 0 and self.IsAvailableShed[n] > 0:
NumDevicesToCall += 1
P_request_perRF = P_req/max(NumDevicesToCall,1) #divide the fleet request by the number of devices that can be called upon
#################################
for rfdg in self.RFs: #loop through all RFs
TsetLast = self.TsetInitial[number]
TairLastB = self.TairInitialB[number]
TfoodLastB = self.TfoodInitialB[number]
TcaseLastB = self.TcaseInitialB[number]
TairLast = self.TairInitial[number]
TfoodLast = self.TfoodInitial[number]
TcaseLast = self.TcaseInitial[number]
if P_req<0 and self.IsAvailableAdd[number] > 0 :
# For step >=1, can use last step temperature...
response = rfdg.FRIDGE(TairLastB, TfoodLastB, TcaseLastB, TairLast, TfoodLast, TcaseLast, TsetLast, self.Tamb[number][self.step], self.R_case[number], self.R_food[number], self.R_infil[number],
self.C_case[number], self.C_food[number], self.C_air[number], P_request_perRF, dt,
self.elementOnB[number], self.elementOn[number], self.lockonB[number], self.lockoffB[number], self.lockon[number], self.lockoff[number],
self.cycle_off_base[number], self.cycle_on_base[number], self.cycle_off_grid[number], self.cycle_on_grid[number])
P_req = P_req - response.Eservice
if P_req >= 0: # all the service request has been met
P_req = 0
elif P_req>0 and self.IsAvailableShed[number] > 0 :
# For step >=1, can use last step temperature...
response = rfdg.FRIDGE(TairLastB, TfoodLastB, TcaseLastB, TairLast, TfoodLast, TcaseLast, TsetLast, self.Tamb[number][self.step], self.R_case[number], self.R_food[number], self.R_infil[number],
self.C_case[number], self.C_food[number], self.C_air[number], P_request_perRF, dt,
self.elementOnB[number], self.elementOn[number], self.lockonB[number], self.lockoffB[number], self.lockon[number], self.lockoff[number],
self.cycle_off_base[number], self.cycle_on_base[number], self.cycle_off_grid[number], self.cycle_on_grid[number])
P_req = P_req - response.Eservice
if P_req <= 0: # all the service request has been met
P_req = 0
# break
else:
response = rfdg.FRIDGE(TairLastB, TfoodLastB, TcaseLastB, TairLast, TfoodLast, TcaseLast, TsetLast, self.Tamb[number][self.step], self.R_case[number], self.R_food[number], self.R_infil[number],
self.C_case[number], self.C_food[number], self.C_air[number], 0, dt,
self.elementOnB[number], self.elementOn[number], self.lockonB[number], self.lockoffB[number], self.lockon[number], self.lockoff[number],
self.cycle_off_base[number], self.cycle_on_base[number], self.cycle_off_grid[number], self.cycle_on_grid[number])
# assign returned parameters to associated lists to be recorded
self.TsetInitial[number] = response.Tset
self.TairInitialB[number] = response.TairB
self.TfoodInitialB[number] = response.TfoodB
self.TcaseInitialB[number] = response.TcaseB
self.elementOnB[number] = response.ElementOnB
self.TairInitial[number] = response.Tair
self.TfoodInitial[number] = response.Tfood
self.TcaseInitial[number] = response.Tcase
self.elementOn[number] = response.ElementOn
self.lockonB[number] = response.lockonB
self.lockoffB[number] = response.lockoffB
self.lockon[number] = response.lockon
self.lockoff[number] = response.lockoff
self.cycle_off_base[number] = response.cycle_off_base
self.cycle_on_base[number] = response.cycle_on_base
self.cycle_off_grid[number] = response.cycle_off_grid
self.cycle_on_grid[number] = response.cycle_on_grid
self.SOC[number] = response.SOC
self.SOCb[number] = response.SOC_b
self.IsAvailableAdd[number] = response.IsAvailableAdd
self.IsAvailableShed[number] = response.IsAvailableShed
self.AvailableCapacityAdd[number] = response.AvailableCapacityAdd
self.AvailableCapacityShed[number] = response.AvailableCapacityShed
self.ServiceCallsAccepted[number] = response.ServiceCallsAccepted
self.ServiceProvided[number] = response.Eservice
p_togrid += response.Eused
# resp.P_togrid_max += response.PusedMax
# resp.P_togrid_min += response.PusedMin
p_service += response.Eservice #Eused/10 #Eservice
p_base += response.Pbase
# FleetResponse.sim_step = response.sim_step
number += 1 # go to next device
if P_req>=0:
service_max += response.AvailableCapacityShed # NOTE THIS ASSUMES THE MAX SERVICE IS LOAD SHED
else:
service_min += response.AvailableCapacityAdd # NOTE THIS ASSUMES THE MIN SERVICE IS LOAD ADD
self.step += 1
# Output Fleet Response
resp = FleetResponse()
# Positive Prequest means reducing power consumption
resp.ts = ts
resp.sim_step = timedelta(seconds=dt)
resp.P_service = []
resp.P_service_max = [] # positive only?
resp.P_service_min = []
resp.P_togrid = []
resp.P_togrid_max = []
resp.P_togrid_min = []
resp.P_forecast = []
resp.P_base = []
resp.E = []
resp.C = []
# resp.P_dot_up = resp.P_togrid_max / ServiceRequest.Timestep.seconds
resp.P_service = p_service
resp.P_service_max = service_max # positive decrease loads
resp.P_service_min = service_min # negative increase loasd
resp.P_base = -p_base
resp.P_togrid = -p_togrid
resp.P_togrid_max = -(p_base-service_max)
resp.P_togrid_min = -(p_base-service_min)
resp.E = 0
resp.C = 0 # response.SOC/(self.numRF)
resp.Q_togrid = 'NA'
resp.Q_service = 'NA'
resp.Q_service_max = 'NA'
resp.Q_service_min = 'NA'
resp.Q_togrid_min = 'NA'
resp.Q_togrid_max = 'NA'
resp.Q_dot_up = 'NA'
resp.Q_dot_down = 'NA'
resp.P_dot_up = 0
resp.P_dot_down = 0
resp.dT_hold_limit = 'NA'
resp.T_restore = 'NA'
resp.Strike_price = 'NA'
resp.SOC_cost = 'NA'
# TotalServiceProvidedPerTimeStep[step] = -1.0*self.P_service # per time step for all hvacs -1.0*
"""
Modify the power and SOC of the different subfeets according
to the frequency droop regulation according to IEEE standard
"""
if self.FW21_Enabled and self.is_autonomous:
power_ac = resp.P_service_max
p_prev = resp.P_togrid
power_fleet = self.frequency_watt(power_ac, p_prev, self.ts, self.location, self.db_UF_subfleet, self.db_OF_subfleet)
power_fleet_step, SOC_step = self.update_soc_due_to_frequency_droop(initSOC, power_fleet, dt)
self.time = t + dt
self.ts = self.ts + timedelta(seconds = dt)
# Restart time if it surpasses 24 hours
if self.time > 24*3600:
self.time = self.time - 24*3600
# Impact Metrics
# Update the metrics
Ttotal = 0
SOCTotal = 0
self.cycle_basee = np.sum(self.cycle_off_base)
self.cycle_basee += np.sum(self.cycle_on_base)
self.cycle_grid = np.sum(self.cycle_off_grid)
self.cycle_grid += np.sum(self.cycle_on_grid)
for numberr in range(self.numRF-1):
if self.TairInitial[numberr]>=self.TsetInitial[numberr] + 3: #assume 2F, self.deadband
self.unmet_hours += 1*self.sim_step/3600.0
self.ave_TairB = np.average(self.TairInitialB)
self.ave_Tair = np.average(self.TairInitial)
self.SOCb_metric = np.average(self.SOCb)
self.SOC_metric = np.average(self.SOC)
self.unmet_hours = self.unmet_hours/self.numRF
self.ratio_P_togrid_P_base = resp.P_togrid/(resp.P_base)
self.energy_impacts += abs(resp.P_service)*(self.sim_step/3600)
return resp