Skip to content
This repository has been archived by the owner on Sep 11, 2024. It is now read-only.

Latest commit

 

History

History
297 lines (219 loc) · 11.1 KB

README.md

File metadata and controls

297 lines (219 loc) · 11.1 KB
Raw

build

ottermaticslib

A library to tabulate information from complex systems with various ways to store data and act as glue code for tough engineering problems.

Installation

pip install git+https://github.com/Ottermatics/ottermaticslib.git

Core Functions

  1. Tabulation Of Complex Systems
  2. Modular Post Processing (dataframes)
  3. Exploratory Analysis (ipython + functions / docs)
  4. Workflows for core engineering problemes (structures + cost, thermal+fluids solve)

MVP Features (WIP)

  1. Tabulation, use attrs.field and system_property to capture y=f(x) [Done]
  2. Dynamic Programing ensures work is only done when new data is available with cached_system_property. [Done]
  3. Quick Calculation provided by direct cached references to attribues and properties [Done]
  4. Solver based on NPSS strategy of balances and integrators [Done]
  5. Reporting to google sheets, csv and excel.

Systems & Analysis

Systems record data from components, and can execute a solver via the run(**parameter_iterables) command. Via a system's run command its state and internal component's & systems state can be altered in an outer product fashion, ie all combinations of inputs will be run. At the start of a run the systems & its components state is recorded and reset by default using Ref instances so that way multiple systems can use the same component. Its possible reference loops may occcur so its generally preferred to create components per system, however for coupled systems this is often desireable to converge on a solution.

By default the system calls a default_solver() method in its execute() function. A solver aims to drive its dependent parameter to zero by changing the independent parameters to zero, however it may adjust multiple parameters to meet multiple targets in more complex applications. For custom System behavior or to invoke custom solvers this method may be overriden.

To use the default solver & constraints

SolverSystem(System):
    sol2 = SOLVER.define("dep", "indep")
    sol2.add_constraint("max", limit_max) #indep should never go above this value (or function)
    sol2.add_constraint("min", 0) #indep should never go below zero

Analysis is a pluggable way to provide different output and calculation from the same system and interacts with plot and table reporters.

Components, Signals & Slots

Component are able to be mounted into multiple Systems via SLOTS.define( ComponentType ). A Component's properties can be updated via SIGNALS in the Systems's solver in the pre_execute and/or the post_execute functions via SIGNAL.define(target, source, mode) where mode can be pre,post or both to update before the System.execute() method.

Iterable Components may be defined on a System via SLOT.define_iterable( <ComponentIter>, wide=True/False) to choose how the system should iterate over the component, wide mode provides all the component attributes and properties in the same row whereas the narrow mode will iterate over each combination of component as though it was input into system.run()

Example Engineering Problems:

These problems demonstrate functionality

Air Filter

run a throttle sweep with filter loss characteristic and fan afinity law based pressure based off of a design point.

from ottermatics.analysis import Analysis
from ottermatics.reporting import CSVReporter,DiskPlotReporter
from ottermatics.properties import system_property
from ottermatics.components import Component
from ottermatics.system import System
from ottermatics.plotting import PLOT
from ottermatics.slots import SLOT
from ottermatics.solver import SOLVER
from ottermatics.signals import SIGNAL
from ottermatics.configuration import otterize
import numpy as np
import os,pathlib
import attrs

@otterize
class Fan(Component):

    n:float = attrs.field(default=1)
    dp_design= attrs.field(default=100)
    w_design = attrs.field(default=2)


    @system_property
    def dP_fan(self) -> float:
        return self.dp_design*(self.n*self.w_design)**2.0
    
@otterize
class Filter(Component):

    w:float = attrs.field(default=0)
    k_loss:float = attrs.field(default=50)

    @system_property
    def dP_filter(self) -> float:
        return self.k_loss*self.w

@otterize
class Airfilter(System):

    throttle:float = attrs.field(default=1)
    w:float = attrs.field(default=1)
    k_parasitic:float = attrs.field(default=0.1)

    fan: Fan = SLOT.define(Fan)
    filt: Filter = SLOT.define(Filter)

    set_fan_n = SIGNAL.define('fan.n','throttle',mode='both')
    set_filter_w = SIGNAL.define('filt.w','w',mode='both')

    flow_solver = SOLVER.define('sum_dP','w')
    flow_solver.add_constraint('min',0)

    flow_curve = PLOT.define(
        "throttle", "w", kind="lineplot", title="Flow Curve"
    )    

    @system_property
    def dP_parasitic(self) -> float:
        return self.k_parasitic * self.w**2.0

    @system_property
    def sum_dP(self) -> float:
        return self.fan.dP_fan - self.dP_parasitic - self.filt.dP_filter


#Run the system
from ottermatics.logging import change_all_log_levels
from matplotlib.pylab import *

fan = Fan()
filt = Filter()
af = Airfilter(fan=fan,filt=filt)

af.run(throttle=list(np.arange(0.1,1.1,0.1)))

df = af.dataframe

fig,(ax,ax2) = subplots(2,1)
ax.plot(df.throttle*100,df.w,'k--',label='flow')
ax2.plot(df.throttle*100,filt.dataframe.dp_filter,label='filter')
ax2.plot(df.throttle*100,df.dp_parasitic,label='parasitic')
ax2.plot(df.throttle*100,fan.dataframe.dp_fan,label='fan')
ax.legend(loc='upper right')
ax.set_title('flow')
ax.grid()
ax2.legend()
ax2.grid()
ax2.set_title(f'pressure')
ax2.set_xlabel(f'throttle%')
Results

air_filter_calc.png

Spring Mass Damper

Overview

Test case results in accurate resonance frequency calculation

@otterize
class SpringMass(System):

    k:float = attrs.field(default=50)
    m:float = attrs.field(default=1)
    g:float = attrs.field(default=9.81)
    u:float = attrs.field(default=0.3)

    a:float = attrs.field(default=0)
    x:float = attrs.field(default=0.0)
    v:float = attrs.field(default=0.0)
    t:float = attrs.field(default=0.0)

    x_neutral:float = attrs.field(default=0.5)
    
    #a is solved for to ensure sumF is zero
    res = SOLVER.define('sumF','a')

	#a is integrated to provide v, similar to v integrated to supply x
    vtx = TRANSIENT.define('v','a')
    xtx = TRANSIENT.define('x','v')

    @system_property
    def dx(self)-> float:
        return self.x_neutral- self.x 

    @system_property
    def Fspring(self)-> float:
        return self.k * self.dx
    
    @system_property
    def Fgrav(self)-> float:
        return self.g * self.m
    
    @system_property
    def Faccel(self)-> float:
        return self.a * self.m
    
    @system_property
    def Ffric(self)->float:
        return self.u*self.v

    @system_property
    def sumF(self) -> float:
        return self.Fspring - self.Fgrav - self.Faccel - self.Ffric


#Run The System, Compare damping `u`=0 & 0.1
sm = SpringMass(x=0.0)
sm.run(dt=0.01,endtime=10,u=[0.0,0.1])

df = sm.dataframe
df.groupby('run_id').plot('time','x')
Results Damping Off

olib_spring_mass_clac.png

Results - Damping On

Reporting & Analysis

Analysis is capable of tabulation as a Component or System and wraps a top level System and will save data for each system interval. Analysis stores several reporters for tables and plots that may be used to store results in multiple locations.

Reporting is supported for tables via dataframes in CSV,Excel and Gsheets (WIP).

For plots reporting is supported in disk storage.

from ottermatics.analysis import Analysis
from ottermatics.reporting import CSVReporter,DiskPlotReporter
from ottermatics.properties import system_property
import numpy as np
import os,pathlib

this_dir = str(pathlib.Path(__file__).parent)
this_dir = os.path.join(this_dir,'airfilter_report')
if not os.path.exists(this_dir):
    os.path.mkdir(this_dir)

csv = CSVReporter(path=this_dir,report_mode='daily')
csv_latest = CSVReporter(path=this_dir,report_mode='single')

plots = DiskPlotReporter(path=this_dir,report_mode='monthly')
plots_latest = DiskPlotReporter(path=this_dir,report_mode='single')


@otterize
class AirfilterAnalysis(Analysis):
    """Does post processing on a system"""
    
    efficiency = attrs.field(defualt=0.95)

    @system_property
    def clean_air_delivery_rate(self) -> float:
        return self.system.w*self.efficiency

    def post_process(self,*run_args,**run_kwargs):
        pass
        #TODO: something custom!

#Air Filter as before
fan = Fan()
filt = Filter()
af = Airfilter(fan=fan,filt=filt)

#Make The Analysis
sa = AirfilterAnalysis(
                    system = af,
                    table_reporters = [csv,csv_latest],
                    plot_reporters = [plots,plots_latest]
                   )

#Run the analysis! Input passed to system
sa.run(throttle=list(np.arange(0.1,1.1,0.1)))

#CSV's & Plots available in ./airfilter_report!

Documentation:

https://ottermatics.github.io/ottermaticslib/build/html/index.html

DataStores

Datastores are a work in progress feature to provide a zero configuration library for storage of tabulated data and report generated artifacts. No garuntee is provided as to their stability yet. Requirements for datasources are attempted upon access of ottermatics.datastores and entering of a CONFIRM prompt.

Environmental Variables

To allow a write-once implement anywhere interface EnvVariable is provided for both open (the default) and secret variables. Allowance for type conversion, and defaults are provided.

The current variable slots in memory are listed by EnvVariable.print_env_vars()

OTTR_DB_HOST                            |SECRETS[OTTR_DB_HOST]                    = localhost
OTTR_DB_NAME                            |SECRETS[OTTR_DB_NAME]                    = 
OTTR_DB_PASS                            |SECRETS[OTTR_DB_PASS]                    = postgres
OTTR_DB_PORT                            |SECRETS[OTTR_DB_PORT]                    = 5432
OTTR_DB_USER                            |SECRETS[OTTR_DB_USER]                    = postgres
OTTR_HOSTNAME                           |SECRETS[OTTR_HOSTNAME]                   = DEATHRAY
OTTR_REPORT_PATH                        |SECRETS[OTTR_REPORT_PATH]                = 
OTTR_SLACK_LOG_WEBHOOK                  |SECRETS[OTTR_SLACK_LOG_WEBHOOK]          = 
SEABORN_CONTEXT                         |SECRETS[SEABORN_CONTEXT]                 = paper
SEABORN_PALETTE                         |SECRETS[SEABORN_PALETTE]                 = deep
SEABORN_THEME                           |SECRETS[SEABORN_THEME]                   = darkgrid