Thursday, April 28, 2022

Notes Day 1 National workshop on electric aircraft and allied technologies - 4

 Biju Prasad B, VSSC ISRO

  1. Has developed fault tolerant EMA for isro
  2. ansys motor cad good for designing motors
  3. Common 4 aviation motors:
    1. magnix 65 - US, 200kg, 640kW, 3020 Nm, 3.2 kW/kg
    2. Siemens SP2 50 kg motor
    3. YASA 750R 37kg axial flux 250kW 200Nm torque 5kW/kg
    4. EMRAX 41 kg 380 kW, outlier power density, may be cheating
  4. Vita technologies make ISRO actuator
  5. 264 kN  stall force, 64 kW power, 270V 110A in 4 coils, with quadruplex windings and quadruplex LVDT
  6. In BLDC, only 2 coils are used at a time
  7. Motor duty cycle for hansa
    1. 68kW for 30S
    2. 46kW for 16.5 mins
    3. 29 kW for 55 mins
    4. 4.1kW for 16.5mins
    5. 17kW for 30 S
  8. PMAC gives 6kW/kg power density, high cost, alll phases used together, sinusoidal commutation gives full utilisation, smooth torque, not like bldc
  9. Axial flux = 10 kW/kg, low inertial, torque is high, stacking of rotor stampings difficult
  10. radial flux = 3kW/kg high inertia, stator construction is simpler, stacking can be wire cut in edm
  11. motors have to satisfy rtca d0 160e qualification reqt for airborne equipment

Dr Karthick sitapaty, Vita technology 

  1. Sine drive is used for high speed motors

  2. Bldc drive is used for simple Motors which are commonly used 

  3. Sine drive is used for machine tool spindles 

  4. In Sin drive LDL Q ratio is important 

  5. Bldc Motors can be converted to PM AC motor 

  6. Cobalt alloy tape material can take up to 2.5 Tesla 

  7. Normal Silicon Steel can take only up to 1.8 Tesla 

  8. When doing simulation for motors there should be at least three layers of cells in the air gap layer 

  9. This is important to get good torque calculation from the finite element model of the 

  10. Skew in the stamping reduces cogging torque in the Motors

  11.  AC drive has to be used with sine wave 

  12. Bldc drive cannot be used sine wave 

  13. AC drive gives 50% more are compared to bldc motor 

  14. For a radial flux motor torque is proportional to square of diameter 

  15. For an axial flux motor torque is proportional to cube of diameter

  16.  2 stator 2  Rotor model will decrease the load on the bearing, as only the differential load comes to the bearing. This is a novel concept being explored by ISRO and Vita Technologies 


  17. Stator is built using a tape of Cobalt alloy 

  18. Tape of Cobalt steel is wounded like a spool of tape

  19. After winding, the slots are milled in the stator. 

  20. Magnet is fit for a trial purposes

  21. Then the tape is unwounded

  22. Then it is annealed through heat treatment

  23. Then an adhesive layer is applied to the Cobalt alloy based tape 

  24. Then the tape is rolled back 

  25. The milled slots may not align again to insert the magnet due to the thickness of adhesive between the layers of the tape 

  26. This requires trial and error.

 


Notes Day 1 National workshop on electric aircraft and allied technologies - 3

 Amitabh Saraf, ADA on FCS

  1. elector hydrostatic actuator =eha
  2. ema have higher bandwidth
  3. magnetic gear unit have inherent torque limiter function, non contact torque transmission
  4. Geared GMR position sensors are v good
  5. 16 kW power required to make all fcs actuators in rta as electric
  6. RCI, NFTDC, VSSC make BLDC electric motors
  7. CPRI makes power drive electronics

Sudharshan Kaarthik  , IIST on motor

  1. Marathon motors and remy motors are common
  2. best for aviation = PMSM with additional reluctance torque, also called PMAC motor, rotor gets heated, problem is how to cool.
  3. Total torque is composed or magnet torque and reluctance torque
  4. Brusa electronik 156 kW, 305Nm torque, PM assisted syn reluctance motor is suitable. 
  5. Magnets in the slot reduces the flux path area but increases the magnet torque. Should compromise
  6. Disk like motor are axial flux syn rm, high torque because of increased number of poles 
  7. motor controllers
    1. unitek bamocar d3
    2. hbc 400400
    3. sd100
    4. pm100
    5. axiom 100+kW
    6. yuneec
    7. wright electric - 2MW continuous power , 1000V DC bus
  8. Hansa needs axial flux, pm assisten syn reluctance motor, 100kW, 20A, 500V DC link
  9. Use silicon carbide switches compared to only silicon. SiC switches have good part load performance. Partload efficiency important when designing with TO power and using at cruise power
  10. multphase machines reduce current per phase and give smooth torque
  11. redundancy is good if one coils fail
  12. SiC have 10 times breakdown electric field strength of Si
  13. Lq/Ld=9 is the most important design parameter of the  motor
  14. field oriented control should be used  as it does independent control of torque and flux
  15. need to model slot harmonics, torque ripple and cogging torque

Notes Day 1 National workshop on electric aircraft and allied technologies - 2

 Youtube video link = https://www.youtube.com/watch?v=2xm_8dY0Vgk




 

 Dr B R Pai

  1. Hybrid hang glider 
  2. weight shift control
  3. 100-200km range
  4. low stall speed of 55 km/hr
  5. Does not need certified engine
  6. 1998 Wankel powered hang glider was installed
  7. Aeros Ant E trike is single seater, powered hang glider, ukraine, not commercial
  8. EMRAX 268 electric motor is good
  9. HBC TMM Controller is suitable
  10. Battery weighs 11 kg, from epsenergy.com
  11. AX4130 couples the engine and motor

Divakar, landing gear, Infosys

  1. Pneumatic actuator not used in a/c
  2. hydromechanical valves are hydraulic actuators controlled by selector valves or servo valves
    1. they have high power density
    2. good for flight safe reliability
  3. electrohydraulic actuators = hydraulic actuators controlled by electrically signalled selector valves, EHSV
  4. Hydraulic actuators reverts to a default damper mode upon failure
  5. EHV = elcetro hydro static actuator - power is generated as per demand, signalled by variable speed brushless motor, typically by a constant displacement pump
  6. EMA = electro mechanical actuator = dc motor - gearing - ball screw to powr a linear actuator rod, precise control and health monitoring possible
  7. EMA ball screw can jam
    1. low power density
    2. limited to secondary flight controls only
    3. require lubrication

J Jayadev Vyas - electric ECS system

  1. Hot day grounded aircraft with pax and hot day cruise are maximum load cases for ecs
  2. 21 kW = estimated heat load of rta, mainly from conduction, convection, metabolic heat
  3. ECS system can be 
    1. 1 wheel turbine = Dornier 228
    2. 2 wheel turbine = saras mk 2
    3. 3 wheel turbine = atr 72
    4. 4 wheel system = rta
  4. Fuel penalty for bleeding hpc air = 1.55 lb/hr for taking 1 lb/min form the hpc outlet
  5. system weight of the electric ecs is higher than normal engine bleed based ecs

ASERDC - HAL Lucknow

  1. Eelctric brakes are 
    1. resistive to oil leakage, 
    2. fire
    3. weight saving
    4. can monitor health
  2. high torque motors required to actuate the ball screw that presses brake disk
  3.  brake temperature = 1000 degree celsius
  4. interface temperature of motor = 150 degree celsius
  5. motor stall torque needs to be high
  6. Frictional locking of shaft required for parking and for braking
  7. Vibhor.asthana@hal-india.com

Notes Day 1 National workshop on electric aircraft and allied technologies - 1

Dr Abhay Pashilkar, BTECH IIT Kharagpur
  1. Passenger traffic to be 520 million bt 2037
  2. Aircraft fleet to be 2500 by 2038
  3. Aircraft movement grew by 9.56% in 2016-2020
  4. India is already 3rd largest domestic aviation market
  5. India set to become 3rd largest overall aviation market
  6. GIFT city to have leasing and finance of airlines through international financial services center 
  7.  Motor to be hallbach array based on ironless motor
  8. battery energy density of 1150 whr/kg required
  9. For PAX less than 19 seater , hydrogen aircrafts are viable.
  10. UK has flyzero concept for zero emission aircraft
  11. ATR72-600 is improved into ATR72-2030 capable of using sustainable aviation fuel
  12. Small aircraft can be powered by Synthetic aviation fuel SAF, hybrid, electric and fuel cell
  13. Distributed propulsion is more effecient for aerodynamics

A C Pankaj on E-HANSA

  1. MTOW 750 kg for JAR VLA aircraft
  2.  Hansa is an all composite low wing monoplane powered by Rotax 912 iSc engine 100 hp
  3. Drag Coeff = 0.0327 + 0.055 Cl^2
  4. TO = 56.8 kW
  5. Climb = 45.9kW
  6. Cruise = 28.8 kW
  7. Descent = 3 kW
  8. Engine fuel = 113.4 kg
  9. Propeller = 6 kg
  10. Chosen motor = Emrax, 500VDC , 5500 RPM, slovenian motor
  11. motor = 12.3 kg
  12. controller = 7.7 kg
  13. Cooler = 6.0 kg
  14. Battery reqd for TO = 72 kg
  15. Battery reqd for Cruise = 55 kg
  16. Fuel enery density = 1250 Whr/kg
  17. Present battery = 272 whr/kg
  18. Endurance  = 52 mn
  19. Range = 126 km
  20. Total energy reqd for a typical mission = 35.6 kW/hr
  21. To carry one kg of H2, 26 kg of composite tank is required.
  22. For small aircraft = battery is best
  23. For medium aircraft = hybric configuration is best
  24. For large aircraft = One fuel cell is better
  25. Safety and reliability of electrical power systems is still unproven
  26. Battery degradation over the life needed to be considered for range calculation

Wednesday, March 30, 2022

#1 Helicopter info collection, for design data base

All data here is fictional/imaginary.

Nothing in this belongs to me.

This is a record of work, for future use only.

IndexNameLinkDisk loadingMain rotor diameterTail rotor diameterMain rotor to tail rotorPowerMTOWCeiling without ground effectEnginePayloadMax ascent speedService ceilingCruise speedEnduranceMax speedFuselage length



kg/m2mmmWkgm
kgm/smm/shrm/sm
1AeroDreams Chi-7https://sites.google.com/site/stingrayslistofrotorcraft/aerodreams-chi-715.9235668861.1NA746004501800Rotax 9122308250036.111111111052.777777787.15

Tuesday, March 22, 2022

Learning Neural networks 3

      -*- coding: utf-8 -*-
"""
@author: Balaji
"""
print('\n 18 March Program for a simple neural network, doing a forward pass with one layer')
print('\n 21 March Added provision for data generation for use in training.')
print('\n 21 March Formating input data as a matrix for faster multiplication with weights')
print('\n 22 March Initialise bias to 0, weights to very small ~0.01 random numbers, put computations inside a class')

import numpy as np
import matplotlib.pyplot as plt

class Layer_dense():
def __init__(self, num_pixels, num_neurons):
np.random.seed(1); #For repeatability
self.weights = np.random.randn(num_neurons ,num_pixels);
self.biases = np.random.randn(num_neurons ,1);
#self.output_beforeActivation = np.zeros((num_neurons,1));
def forward(self, inputs):
self.output_beforeActivation = np.dot(self.weights,inputs)+self.biases;
def applyReluActivation (self):
self.output_afterActivation = np.maximum(0,self.output_beforeActivation);

num_neurons = 4;
num_pixels = 5;
num_images = 10;

# Input data set properties
slope = np.tan(np.deg2rad(30));
intercept = 50;

print('\n Creating input data:')
np.random.seed(1); #For repeatability
inputs = np.zeros((num_pixels,num_images));
x = np.arange(0,num_pixels);
x = x.reshape((num_pixels,1));
for dataSet_iter in range(num_images):
inputs[:,dataSet_iter:dataSet_iter+1] = (slope*x+intercept) + (intercept*0.1)*(np.random.randn(x.size,1));

print('\n Plotting input data set :')
plt.plot(x,inputs,'o')
plt.xlabel('Pixel number')
plt.ylabel('Pixel value')
plt.title('Input data set')
plt.legend()
plt.show()

print('\n Create a dense layer, fully connected, of the class Layer_dense')
firstLayer = Layer_dense(num_pixels = num_pixels, num_neurons = num_neurons);
firstLayer.forward(inputs=inputs);
firstLayer.applyReluActivation();

print('\n Result = ')
print('\n Each column is result of the network per image, after activation')
print(firstLayer.output_afterActivation)

 

Monday, March 21, 2022

Learning neural networks 2

# -*- coding: utf-8 -*-
"""
@author: Balaji
"""
print('\n 18 March Program for a simple neural network, doing a forward pass with one layer')
print('\n 21 March Added provision for data generation for use in training.')
print('\n 21 March Formating input data as a matrix for faster multiplication with weights')

import numpy as np
import  matplotlib.pyplot as plt
num_neurons     = 4;
num_pixels      = 2;
num_images      = 3;
np.random.seed(1); #For repeatability
# Input data set properties
slope           = np.tan(np.deg2rad(30));
intercept       = 50;

print('\n Creating input data:')
inputs  = np.zeros((num_images,num_pixels));
x = np.arange(0,num_pixels);
for dataSet_iter in range(num_images):
    inputs[dataSet_iter,:] = (slope*x+intercept)+ intercept*0.2*(np.random.randn(1,x.size) );

print('\n Plotting input data set :')    
plt.plot(x,inputs.transpose(),'o')
plt.xlabel('Pixel number')
plt.ylabel('Pixel value')
plt.title('Input data set')
plt.legend()
plt.show()

print('\n Each row is a data set')
weights = np.random.randn(num_neurons ,num_pixels);
bias    = np.random.randn(num_neurons ,1);
results = np.zeros_like(bias)

print('\n Input = ')
print(inputs)

print('\n Weights = ')
print(weights)

print('\n Bias = ')
print(bias)

# Output computation in matrix format
results = np.dot(weights,inputs.transpose())+bias;

# Output  computation in loop method
for input_iter in inputs:
    for weight_iter,bias_iter,result_iter in zip(weights,bias,results):
        print('\n Input data set:')
        print(input_iter);
        print('\n is multiplied with weights')
        print(weight_iter)
        print('\n and added with bias')
        print(bias_iter)
        print('\n to give the result of neuron as')
        result_iter = input_iter.dot(weight_iter) + bias_iter;
        print(result_iter)