{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 0,
     "link": "[](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html)",
     "section": ""
    }
   },
   "source": [
    "<!--NOTEBOOK_HEADER-->\n",
    "*This notebook contains material from [CBE30338](https://jckantor.github.io/CBE30338);\n",
    "content is available [on Github](https://github.com/jckantor/CBE30338.git).*\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 0,
     "link": "[](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html)",
     "section": ""
    }
   },
   "source": [
    "<!--NAVIGATION-->\n",
    "< [4.0 PID Control](https://jckantor.github.io/CBE30338/04.00-PID_Control.html) | [Contents](toc.html) | [Tag Index](tag_index.html) | [4.2 PID Control with Setpoint Weighting](https://jckantor.github.io/CBE30338/04.02-PID_Control_with_Setpoint_Weighting.html) ><p><a href=\"https://colab.research.google.com/github/jckantor/CBE30338/blob/master/docs/04.01-Implementing_PID_Control_with_Python_Yield_Statement.ipynb\"> <img align=\"left\" src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open in Colab\" title=\"Open in Google Colaboratory\"></a><p><a href=\"https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.ipynb\"> <img align=\"left\" src=\"https://img.shields.io/badge/Github-Download-blue.svg\" alt=\"Download\" title=\"Download Notebook\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 1,
     "link": "[4.1 Implementing PID Controllers with Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1-Implementing-PID-Controllers-with-Python-Yield-Statement)",
     "section": "4.1 Implementing PID Controllers with Python Yield Statement"
    }
   },
   "source": [
    "# 4.1 Implementing PID Controllers with Python Yield Statement\n",
    "\n",
    "Up to this point we have been implementing simple control strategies where the manipulated variable depends only on current values of the process variable and setpoint. Relay control, for example, can be expressed mathematically as\n",
    "\n",
    "$$MV = Kp (SP - PV)$$\n",
    "\n",
    "and implemented in Python as\n",
    "\n",
    "    def proportional(PV, SP):\n",
    "        MV = Kp * (SP - PV)\n",
    "        return MV\n",
    "\n",
    "Python functions carry no persistent memory from one use to the next, which is fine if the control algorithm requires no knowledge of the past.\n",
    "\n",
    "The problem, however, is that PID control requires use of past measurements. Proportional-integral control, for example, tracks the cumulative sum of the differences between setpoint and the process variable. Because a Python function disappears completely after the return statement, the value of the cumulative sum must be stored somewhere else in code. The coding problem is to figure out how and where to store that information between calls to the algorithm. We seek coding methods that encapsulate the state of control algorithms.\n",
    "\n",
    "There are several ways persistence can be provided in Python (roughly in order of increasing capability and complexity):\n",
    "\n",
    "* Generators written using the Python `yield` statement\n",
    "* Classes\n",
    "* Asynchronous programming with co-routines\n",
    "* Threads\n",
    "\n",
    "Classes, in particular, are a tool for encapsulating data and functionality within a single software entity. Classes are widely used in Python programming, and should eventually become part of every Python programmer's toolkit.\n",
    "\n",
    "As we demonstrate below, however, the Python `yield` statement is often sufficient to write self-contained implementations of control algorithm. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "## 4.1.1 Python Yield Statement\n",
    "\n",
    "The cells below provide a very brief introduction to Python `yield` statement. More information can be found [here](https://jeffknupp.com/blog/2013/04/07/improve-your-python-yield-and-generators-explained/)\n",
    "\n",
    "A function incorporating a `yield` statement creates a persistent object called a generator.  Like the `return` statement, `yield` says to return a value and control to the calling routine. Unlike the `return` statement, however, the generator goes into a suspended state following the `yield`. On the next use it will pick up execution right where execution left off.\n",
    "\n",
    "Here's an example produces a generator for four numbers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": [
    "def numbers_gen():\n",
    "    yield 0\n",
    "    yield 1\n",
    "    yield 2.7182818\n",
    "    yield 3.1415926"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "To use this, first create the actual generator object that will do the work. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": [
    "numbers = numbers_gen()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "There are several ways to get values from a generator. One way is using the `next()` function which executes generator until the next `yield` statement is encountered, then return the value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2.7182818\n",
      "3.1415926\n"
     ]
    }
   ],
   "source": [
    "print(next(numbers))\n",
    "print(next(numbers))\n",
    "print(next(numbers))\n",
    "print(next(numbers))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "The generator object `numbers` has a `send` method for sending information to the generator and causing it to execute until the next `yield`.  For this case all we need to send is `None`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2.7182818\n",
      "3.1415926\n"
     ]
    }
   ],
   "source": [
    "numbers = numbers_gen()\n",
    "print(numbers.send(None))\n",
    "print(numbers.send(None))\n",
    "print(numbers.send(None))\n",
    "print(numbers.send(None))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "The `send` method provides two-way communication with a generator. Here's an example of how that works."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": [
    "def texter_gen():\n",
    "    a = yield \"Started\"\n",
    "    b = yield a\n",
    "    yield b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Started\n",
      "Hello, World\n",
      "Go Irish\n"
     ]
    }
   ],
   "source": [
    "texter = texter_gen()\n",
    "print(texter.send(None))\n",
    "print(texter.send(\"Hello, World\"))\n",
    "print(texter.send(\"Go Irish\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "There's a subtle detail here that easy to miss. The first `send` starts the generator which then executes up to the first yield statement. At that point execution halts and the message \"Started\" is sent to the calling routine. The second `send` causes execution to pick up again and puts the message \"Hello, World\" into variable `a`.  \n",
    "\n",
    "An important point is that the first `send` to a generator must always be `None`.\n",
    "\n",
    "In the next example `yield` is placed inside an infinite loop. This function creates generators that return the square of the number sent."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": [
    "# A function that will create generators\n",
    "def our_numbers():\n",
    "    n = 0\n",
    "    while True:\n",
    "        n = yield n*n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "16\n",
      "144\n",
      "1764\n"
     ]
    }
   ],
   "source": [
    "# create a generator\n",
    "numbers = our_numbers()\n",
    "\n",
    "# start the generator\n",
    "numbers.send(None)        \n",
    "\n",
    "# send values to the generator and print the results\n",
    "print(numbers.send(4))\n",
    "print(numbers.send(12))\n",
    "print(numbers.send(42))\n",
    "\n",
    "# how to remove a generator when finished using it.\n",
    "numbers.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "Let's now consider how generators may be used for process control applications. The next cell defines a function that will create generators that perform as proportional controllers with specified gain $K_p$ and setpoint $SP$. When sent a current value of the process variable `PV`, the controllers will return a value for the manipulated variable `MV`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [],
   "source": [
    "def proportional(Kp, SP):\n",
    "    \"\"\"Creates proportional controllers with specified gain and setpoint.\"\"\"\n",
    "    MV = 0\n",
    "    while True:\n",
    "        PV = yield MV\n",
    "        MV = Kp * (SP - PV)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "Let's create and initialize two controllers with the same setpoint but different gains."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "controller1 = proportional(10, 40)\n",
    "controller1.send(None)\n",
    "\n",
    "controller2 = proportional(1, 40)\n",
    "controller2.send(None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "Let's see how these controllers would respond to a PV value of 35."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Controller 1: MV =  50\n",
      "Controller 2: MV =  5\n"
     ]
    }
   ],
   "source": [
    "PV = 35\n",
    "\n",
    "print(\"Controller 1: MV = \", controller1.send(PV))\n",
    "print(\"Controller 2: MV = \", controller2.send(PV))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.1 Python Yield Statement](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.1-Python-Yield-Statement)",
     "section": "4.1.1 Python Yield Statement"
    }
   },
   "source": [
    "This is a important feature from a coding and maintenance perspective. We need to create and maintain only one copy of the control algorithm."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.2 PID Control Implementation](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.2-PID-Control-Implementation)",
     "section": "4.1.2 PID Control Implementation"
    }
   },
   "source": [
    "## 4.1.2 PID Control Implementation\n",
    "\n",
    "A simple form of Proportional-Integral control is an example of a controller with internal state. In velocity form,\n",
    "\n",
    "\\begin{align}\n",
    "MV_k & = \\overline{MV} + K_P e_k  + K_I \\sum_{k'=0}^k e_{k'}(t_k - t_{k-1}) + K_D \\frac{e_k - e_{k-1}}{t_k - t_{k-1}}\n",
    "\\end{align}\n",
    "\n",
    "where $e_k$ is the difference between setpoint and measured process variable\n",
    "\n",
    "\\begin{align}\n",
    "e_k = SP_k - PV_k\n",
    "\\end{align}\n",
    "\n",
    "$K_P$, $K_I$, and $K_D$ are control constants, and $t_k$ is the sampling time. \n",
    "\n",
    "The following cell provides a direct implementation of this algorithm as a Python generator."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.2 PID Control Implementation](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.2-PID-Control-Implementation)",
     "section": "4.1.2 PID Control Implementation"
    }
   },
   "outputs": [],
   "source": [
    "def PID(Kp, Ki, Kd, MV_bar=0):\n",
    "    # initialize stored data\n",
    "    e_prev = 0\n",
    "    t_prev = -100\n",
    "    I = 0\n",
    "    \n",
    "    # initial control\n",
    "    MV = MV_bar\n",
    "    \n",
    "    while True:\n",
    "        # yield MV, wait for new t, PV, SP\n",
    "        t, PV, SP = yield MV\n",
    "        \n",
    "        # PID calculations\n",
    "        e = SP - PV\n",
    "        \n",
    "        P = Kp*e\n",
    "        I = I + Ki*e*(t - t_prev)\n",
    "        D = Kd*(e - e_prev)/(t - t_prev)\n",
    "        \n",
    "        MV = MV_bar + P + I + D\n",
    "        \n",
    "        # update stored data for next iteration\n",
    "        e_prev = e\n",
    "        t_prev = t"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.3 Simulation](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.3-Simulation)",
     "section": "4.1.3 Simulation"
    }
   },
   "source": [
    "## 4.1.3 Simulation\n",
    "\n",
    "Let's see how well this PID implementation works. We'll perform the simulation with a setpoint that starts at room temperature, then rises to 50°C at $t = 50$ seconds. The data historian will record values of the setpoint, process variable, computed manipulated variable, and the actual value of the manipulated variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.3 Simulation](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.3-Simulation)",
     "section": "4.1.3 Simulation"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmsAAAHYCAYAAAAS+q+iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAMTQAADE0B0s6tTgAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzs3X2UXHd95/nPt0r9qI5Qj9VWgmQjzJpkccC2kLGsMBAvDNJkmWyyIgq2tRGeKHYcx5yMTBLiISfZTLJhkkWbg/F4cQS2hGwTJ9LODoGsWIKJIQZbAtnmIcFORMdYgGn5tNzbUndXd9V3/6i6rVt1q6qrux7urVvv1zk+7rrVXf3rn6qrPv39PZm7CwAAAMmUibsBAAAAqI2wBgAAkGCENQAAgAQjrAEAACQYYQ0AACDBCGsAAAAJRlgDAABIMMIaAABAghHWAAAAEoywBgAAkGCENQAAgARbFec3N7NxSbOl/yTpj939L8zsckkHJa2TdFbSe9z9W0s9Xl9fn69fv75dze1Kc3NzGhgYiLsZiUKflKM/ouiTKPqkHP0RRZ9EnT59OufuTXdKrGGt5F3u/o2Kax+VdJ+7P2Bm75L0MUnXLfVAa9eu1QsvvNCONnatY8eOafv27XE3I1Hok3L0RxR9EkWflKM/ouiTKDObaMXjJCGslTGziyVtlvSO0qUjkj5iZpvcfTy2hi1DvuCaPJ+LuxmSpKmc68z0XNzNSBT6pBz9EUWfRNEn5eiPKPqkfczd4/vmxWHQl1WcO/eEpN+RdKmkT7j760Kf96Sk97n7Y/Ueb2xszCcmWhJiVyxfcN3+4Nd0+uxMrO0IvDz1sl6x5hVxNyNR6JNy9EcUfRJFn5SjP6Lok6i/fu+/Pu3uG5t9nLgra29x9+fNrE/SH6o4T+13JVUmSKv2xWa2T9K+4Pbw8LCOHTvWrrY2ZCrn+tbz+VjbEOYF18tTL8fdjEShT8rRH1H0SRR9Uo7+iKJP2ifWylqYmf2YpGclvUbSc5IucvcFMzNJ35e0dalh0CRU1s5Mz+nm+49LkvbvulKjq/tjbc+jjz6q66+/PtY2JA19Uo7+iKJPouiTcvRHFH0SNfYjg91dWTOz1ZL63P1s6dINkk66+w/N7KSk3ZIekLRT0ni3zFcLG13dr3Uj8a6MWdNvsbchaeiTcvRHFH0SRZ+Uoz+i6JP2iXMYdL2kI2aWVXGY85SkXyrdd6ukB8zsLklTkvbE00QAAIB4xRbW3P2UpKtr3PdtNbBVBwAAQNpxggEAAECCEdYAAAASjLAGAACQYIQ1AACABCOsAQAAJBhhDQAAIMEIawAAAAlGWAMAAEgwwhoAAECCEdYAAAASjLAGAACQYIQ1AACABCOsAQAAJBhhDQAAIMEIawAAAAlGWAMAAEgwwhoAAECCEdYAAAASjLAGAACQYIQ1AACABCOsAQAAJBhhDQAAIMEIawAAAAlGWAMAAEiwRIQ1M/s9M3Mz+8nS7cvN7HEze9bMnjSz18XdRgAAgDjEHtbMbLOkrZKeD13+qKT73P21kv5E0sfiaBsAAEDcYg1rZjYg6R5JvybJS9culrRZ0uHSpx2R9Goz2xRDEwEAAGIVd2XtDyQddvfvhK5dIul77r4gSe7uKlbdLo2hfQAAALGyYhaK4RubXSfpjyS9zd3dzMYlvVPSgKRD7n5F6HOPS7rT3R+reIx9kvYFt4eHhzccPXq0E82vaSrnuvuZvCTpjjdktabfYm3P7OysBgcHY21D0tAn5eiPKPokij4pR39E0SdRO3bsOO3uG5t9nFWtaMwKvVXST0j6jplJ0kZJxyTdJWmjma1y9wUr3nmJyue0SZLcfb+k/cHtsbEx3759eyfaXtOZ6TkdGj8uSbr++mu0bmQg1vYcO3ZMcfdJ0tAn5eiPKPokij4pR39E0SftE9swqLt/0N1f6e6b3H2TpBckbXf3g5JOStpd+tSdksbdfTyelgIAAMQnzspaPbdKesDM7pI0JWlPzO0BAACIRWLCWqm6Fnz8bUnXxdcaAACAZIh7NSgAAADqIKwBAAAkGGENAAAgwQhrAAAACUZYAwAASLCmV4Oa2f8o6cclfdXd/675JgEAACDQVGXNzP6TpLslXSvpITO7pSWtAgAAgKTmK2s7JV3t7i+b2UZJRyTd13yzAAAAIDU/Z23W3V+WJHd/QVJf800CAABAoNnK2ivM7GdCt9eEb7v7Z5p8fAAAgJ7WbFj7rqTfrLj9PkkmySUR1gAAAJrQbFi73t09uGFmr5H0s5L+yd0/1eRjAwAA9Lxm56x91syukiQze6Wk45LeIelPzOy3m20cAABAr2s2rG1w96dKH98o6e/c/d9K2ibppiYfGwAAoOc1vRo09PE2leaoufukpIUmHxsAAKDnNRvWCma20cxWS3qrpPAJBsNNPjYAAEDPa3aBwf8m6auS5iU96u7PSpKZbZM03uRjAwAA9Lymwpq7HzWzv5f0Y5KeDt01LomjpwAAAJrU9EHu7v6ipBcrrn2v2ccFAABA83PWAAAA0EaENQAAgAQjrAEAACQYYQ0AACDBCGsAAAAJFmtYM7PPmtkzZvaUmX0xdM7o5Wb2uJk9a2ZPmtnr4mwnAABAXOKurO1y9ze4+1WSPiTp46XrH5V0n7u/VtKfSPpYXA0EAACIU6xhzd3Phm6+QsXjqy6WtFnS4dL1I5JebWabOts6AACA+DW9KW6zzOyQpOtLN3dIukTS99x9QZLc3c3seUmXiiOsAABAjzF3j7sNkiQz2yPpFyX9rqRD7n5F6L7jku5098cqvmafpH3B7eHh4Q1Hjx7tUIurm8q57n4mL0m64w1Zrem3WNszOzurwcHBWNuQNPRJOfojij6Jok/K0R9R9EnUjh07Trv7xmYfJzFhTZLMbEbSJknPSrrI3RfMzCR9X9JWdx+v9/VjY2M+MTHR9nbWc2Z6Tjfff1ySdP/N12jdyECs7Tl27Ji2b98eaxuShj4pR39E0SdR9Ek5+iOKPokys5aEtdjmrJnZGjN7Zej2z0t6SdIPJZ2UtLt0105J40sFNQAAgDSKc87aKyQdMbMhSQVJE5LeWZqjdqukB8zsLklTkvbE2E4AAIDYxBbW3P27kt5U475vS7qusy0CAABInrj3WQMAAEAdhDUAAIAEI6wBAAAkGGENAAAgwQhrAAAACUZYAwAASDDCGgAAQIIR1gAAABKMsAYAAJBghDUAAIAEI6wBAAAkGGENAAAgwQhrAAAACUZYAwAASDDCGgAAQIIR1gAAABKMsAYAAJBghDUAAIAEI6wBAAAkGGENAAAgwQhrAAAACUZYAwAASDDCGgAAQIKtirsBrVRw15npuVjbMHkuF+v3BwAA6RJbWDOzQUmflPQ6Secl/UDSr7r7uJldLOmQpNdImitd/9JSjzk9L918//E2thoAAKCz4h4GvU/Sj7v7VZL+unRbkj4o6SvufrmkmyU9aGZdVQXcsHZIo8P9cTcDAAB0udgCkLvPSvpM6NJXJP1G6eNdkl5d+rzjZvaipDdL+kK9xxzpk+6/+ZrWN3YFRof7lc1Y3M0AAABdLknVqvdK+pSZXSQp4+4TofvGJV261ANkzLRuZKBNzQMAAOg8c/e42yAzu0vSv5P0NklDkp5399Wh+/9S0qfc/VDF1+2TtC+4PTw8vOHo0aOdaXSXmJ2d1eDgYNzNSBT6pBz9EUWfRNEn5eiPKPokaseOHafdfWOzjxN7WDOz90l6t6S3u/vZ0rVzkjYF1TUze1LSb7n7F+o91tjYmE9MTNT7lJ5z7Ngxbd++Pe5mJAp9Uo7+iKJPouiTcvRHFH0SZWbdH9ZKlbGbVAxqk6HrD0gad/ffN7NrJB2RdJm7L9R7vL6+Pl+/fn07m9x15ubmNDDA0HAYfVKO/oiiT6Lok3L0RxR9EnX69Gm5e9MT2GMLa2a2UdJ3JZ2S9P+VLs+5+7Vmtl7SJ1RcZJCT9Gvu/ndLPSaVtSj+0omiT8rRH1H0SRR9Uo7+iKJPosws7+5Nrw+IczXoC5Kqpk13f1HSOzrbIiB58gXX5Pn6Gy2z8hgA0i1Jq0EB6EJAyxdcdzx0UjPz+bqfP9SX1d03Xl0W2AhwAJAehDUgJtWqZo0GtLCZ+bz2HjxRdq1agJMIcQDQYdOteBDCGtBh+ULxDNtGQ1mt4FUv2FULcMFjHd57rfpXxX14CQD0BMIakHSV1bOVDG3Wq4Y9fMvWZT3+zHxeO+99XAf2bFl8bABAshHWgDbJLRS0+8ATdYNZs8OV2Uz01I7KACcVQ1y40hZ8PNSX1c5XFit9y/m+AIDOIawBbZBbKGjnvY/XvD8IaetGBloejqoFOEk6ctu2SHicmc/rnq/ndfhfjkuSNqwd0j03bSawAUCCENaAFggPd1ZWsSQtDjsG4qhg9a/KLFbdag2Xnj47ozPTc1q/hiNjACApCGtAExpZLHDktm2JmdAfrroFwe3RRx/VVW+6UvseeVpScYj0wJ4tban6AQCWL9awZmYDkj4kabuKJxWcdPfdZna5pIOS1kk6K+k97v6t+FoKXNDoPmhJX3kZBLc1/abLxkY01Jdd/Fn2HjzR8CIHAEB7xV1Z+6CkgqTXurub2Y+Vrn9U0n3u/oCZvUvSxyRdF1cjgUC9RQOViwW6KeBkM6bDe68t+9nC23+0c44dAKC+2MKama2WdLOkjV46oNTdv29mF0varAvHTR2R9BEz2+Tu47E0Fj0tXEmrtXdZGoJMMKet2rBuENzS8rMCQDeJs7L2GkkvSfqAmb1d0oyk31dx2PN77r4gSaWK2/OSLpU0Hk9T0avqVdLCe5WlJbhkM6b1awZrLkQIQhurRgGgc6xU1Or8NzZ7o6QTkva4+yEzu1LS5yS9U9LH3f2K0Ocel3Snuz9W8Rj7JO0Lbg8PD284evRoR9rfLWZnZzU4yMq+sHp9UnDX9HzwsXTP16MhrT8j/YerslqVkqCy1HOk4K6pnPTn38wrV7hw/eb/PquRvuLHI31SxtLRHxK/N9XQJ+Xojyj6JGrHjh2n3X1js48TZ1hbJ+lFSf3uni9de1LS/y7pzyVd5O4LZmaSvi9p61LDoGNjYz4xMdHehneZY8eOafv27XE3I1Gq9UkjqzrTWEmTGn+O5AuuUxPTi6tGw9JWaeP3Joo+KUd/RNEnUWbWkrAW2zCou58xs79VcSXoZ8zsVZJeLemLkk5K2i3pAUk7JY0zXw3tstRJA0lf1dkp2Uxx1eiGtUM6fXam7L7TZ2c0eT5XdTNeAEBz4l4N+quSPm5m/1lSXtItpUUGt0p6wMzukjQlaU+cjUT61Fs00M2rOtstmzHdc9PmxQ2AJ8/lFittk+eK1+gvAGitWMOau5+S9NNVrn9bbNWBNim46/YHvxapDkliM9gG1DrOKghtaRsSBYC49fa4DnpGMCftzPScfnBekaA21JfVkdu2af2aQULGMowO92vD2qGya6fPzujUxLTyhXjmwwJA2sQ9DAq0Xb5QXkl7eSqvV6wp3rd/15UaXd3P0N0KhYdFw0Oi+x55mhMQAKBFCGtIvTPTc1WHPDesHdJlYyOEiCYFw6JBlS3o6/AJCAyNAsDKEdaQSrUWEOzfdaWeevLvdf3111DtabGgylZtC5TTZ2d0ZnpO69ewBxMALBdhDamx1AHrQ31ZXTY2olP91SfIo3mVJyCEh0b3HjzBAg4AWAHCGlKhcl5apWCvNEJCZ4SHRof6sovBOThflH3rAKBxhDV0vWBn/WorPJngHq9sxnR477Vlmw7PzOe1897HqbKhY4Kqe6uMDve37LGARhDW0LVqHRHFCs9k6V+V0cO3bNWZ6bmy+YNBle3uG68mtGHZGg1gtaZFNGOoL6udryy+/jSC1yI0i7CGrlTriChWeCZTMJftyG3bIlW2vQdPsFoUDWvkHN92m5nP656v53X4X4439PmcioJmEdbQNRo5IooKTbKFq2zhN9tgI12CNqTaVbNmqmSVgWml7VrJ9w9vY9NoWwh0CCOsoSvUW0DA3KfuEl4xempiOrKRLosPelutqnk1ywlgrQo/wUrnRx99VNdff03dz60V7irDWzVU4xBGWEOiBX9hT57LVV1AwBt798pmTJeNjUQ20g0WH7AwJL3qVc6WCjFSvJX0YKXzmga3AArCnbS8ylwj1Th+P3oHYQ2JVesvbBYQpEd4I93KxQcSgTyNllM5C0J7pW763Q/CXSAc3qpZTjWO6lvvIKwhkXILBe289/HIdRYQpE+txQfShUrbkdu2Edi6WL35ptWkOaRXhrdqGq3GLWcuHEGuuxHWkDj5gmv3gSfKrjEsln7B4oNqb+rsy9ZdwsOc9cJGGipn7bBUNW4lc+HYd7K7EdaQGOH5aeEXIaoqvSP8JnXktm1l1VVOP+gOjQxz8u+4PNWqccudCxcOcgyfdh/CGmJXb9+kA3u28ILeo/pXZaruy8YChGSprKJR2emMRufCVQtyy91KhH+v5Wn1iRkSYQ0xq/dX+Ia1Qxy43uPqnX4gic10Y9TI5rThYU7e8Nur3ly48PSClWwlwkrU2iqDWTtOzJAIa4hRtUUE/PWNSvUWIASb6bI6uDPCCwXqvSExzJks4SC3kq1EGlmJKklTudpHcCX593OllbB2BbNqCGvouOAv8spffiaQo57wAoTJc7myzXQlTrFop3zBdXbOdcN9X6n6xsQcqO6xnK1ElrMSVZJensrr0Hj1I7iWGmpdM9inqdn5xdsrfQ4tN3h1MnA1g7CGjqp1EgGLCNCI4I1mdLi/bDNdiXNGW62yivaDl/J6xZryNzQCcvdbaiuRRleiLqWRUxvCVnI8WFzBq1pbg/A59t7WfA/CGjomX3Cdmpgue4NluAQrEWymW21IjnNGV66RYU6mKvSWpVaihlU7gqtT4a5ZzZwdW+v3oJVzrglraLtaE5H377qSN1SsWOU8nGrnjBIq6mt0PzRJ6s8wVQFFtapxtY7ganSodagvqz9791X6jU8+1VR1bCXBK+mvEYQ1tFWt1Z6cRIBWqnXOKMdW1dbofmjBm97xLz2q9WsGO9hCpEWjQ61BYFrqSK6lJD14rQRhDW0TnEQQfjNgjgvaJXzOaLV9pdif7YJqv5uBWhXJjPVuf6G9KsNcI0dy9RrCGlqu1kkEDKGg3YJtPmodW9WLlbZqq+Oq/W6ykhNILsIaWqrW0MqBPVsYQkHHVB5bVeuA+LRW2hrdD03idxPoBm0Ja2Z20N33tOOxkUy19k6TOIkA8ap3QHytsxIL7rG0tVmNnCoQxu8m0B3aVVm7vk2PiwSqtXdaWqsW6D6NVNrCIW5mOq+3vb2Q+GHSRldz1lodx+8m0B1WHNbM7Ie17pK0dqWPi+7C3mnoNpWVtmoBJ1dQ2TCpFH+wWe4ZhGxdAqRHM5U1k/Q2SS9Xuf73TTwuukS1ihp7p6Eb1DsrsdowqdT5w6yXswdaGCuugfRpJqx9VdJF7v5M5R1m9oMmHhcJF17tGQ5q7J2GblS5TcCR27bpZ/70byKf1+hh1rVUnn1Yz0rCWVIqgABar5mwtlNS1Vced39jE4+LBKs1P42KGtKif1VGd16d1TVvLh6bs9zDrNup09U9AMnQTFg74O43tKwl6ApnpuciQY2KGtImY+XVtlYdZr1cVM0ASM2FtZ9oVSPM7Pck/b6k17v7N8zsckkHJa2TdFbSe9z9W636fli+altz7N91pUZX9/MGgtRbzmHWlSrPPlzOmYX8bgGQmgtrLdmIyMw2S9oq6fnQ5Y9Kus/dHzCzd0n6mKTrWvH9sHzVhj6H+rJU09DTlnMkTuXZhwCwHM3srfB6M/thlf8m6mzrUcbMBiTdI+nXVAp/ZnaxpM2SDpc+7YikV5vZpibaiiZUDn0GW3PwpgM0Jgh2/M4AWIlmKmvPSvqZJr//H0g67O7fsQuHBF8i6XvuviBJ7u5m9rykSyWNN/n9sAy1hj6pqAEA0DnmKzxWxcy+5u6bV/yNza6T9EeS3lYKZOOS3ilpQNIhd78i9LnHJd3p7o9VPMY+SfuC28PDwxuOHj260ial0uzsrAYHl3/uX8Fd932zoJdmLzw/+jPSnVdnlbHuDmor7ZO0oj+i6JMo+qQc/RFFn0Tt2LHjtLtvbPZxmt0UtxlvVXGRQlBV2yjpmKS7JG00s1XuvmDFOy9R+Zw2SZK775e0P7g9Njbm27dvb7JZ6XLs2DEtt0+CUwkWnntar+gvXkvTqQQr6ZM0oz+i6JMo+qQc/RFFn7TPisOau1/dzDd29w9K+mBwO6islVaD3ixpt6QHVNzPbdzdx5v5flharUOgGfoEACA+7TrIvVm3SnrAzO6SNCVpT8ztSb1am92yhxoAAPFKTFhz902hj78tturoqGorPjlfEACA+CUmrCE+uYUCKz4BAEio7p8tjqbkC67dB55YvM1mtwAAJAthrYcFqz7DiwnY7BYAgGRhGLRHVVtQcGDPllRszQEAQJoQ1npMvuCaPJ/T5LlcWVDbsHao4XMOAQBA5xDWekhuoaDdB54oG/aUWFAAAECSEdZ6RG6hoJ33Ph65zj5qAAAkG2GtB1Su+JSK89OyGdPocD9BDQCABCOs9YDJ87myoc8jt21jIQEAAF2Cd+wUyxdcUznX5Lnc4jVWfAIA0F2orKVUsDXHt57P6xXjTy9eZ8gTAIDuQoklpSrP+pSKiwlGh/tjahEAAFgJKmspky+4zkzPRc76HF3dz2ICAAC6EGEtRarto9afEVtzAADQxQhrKRFszxEOakN9Wf3qVVmCGgAAXYywlgLVDmQ/sGeL1o0M6HP/72djbBkAAGgWYa3L1TqQff2awRhbBQAAWoXVoF1u8jwHsgMAkGZU1rpYvlC+4S0HsgMAkD6EtS4UbM9xx0Mny+apja5maw4AANKGsNZlqm3PIbHhLQAAaUVY6yK1tue4+8artW5kgKoaAAApRFjrAvmCa/J8TpPnclW35yCkAQCQXoS1hKu2NYfE9hwAAPQKtu5IuFoHsrM9BwAAvYHKWkJxIDsAAJAIa4lUbehzqC/LHmoAAPQghkETJjjnszKoHd57LUENAIAeRGUtQapV1DiVAACA3kZYS4Dw1hyV53wS1AAA6G2EtZjV2pqDihoAAJAIa7GpVU2TqKgBAIALCGsxqHW+J1tzAACASoS1DgiqaMHH4b3TAlTTAABANbGFNTMblPRJSa+TdF7SDyT9qruPm9nFkg5Jeo2kudL1LzX62OFwtFyjw/2StPj1awb7NDU7v6LHCtpyx0MnI1W0wIE9W5TNGNU0AABQVdyVtfsk/Y27u5n9eun2OyR9UNJX3H2HmV0j6a/M7DXuvlDvwQruenFqtm44Wkr/quLWc7mFwoq+vlHB3mnB9wMAAKgmtrDm7rOSPhO69BVJv1H6eJekV5c+77iZvSjpzZK+UO8xp+dVdYhxOdoV0ob6srr7xqsXq2dU0gAAQCPirqyFvVfSp8zsIkkZd58I3Tcu6dLlPFhlOFpK5XBluMK23MeqhnAGAABWwtw97jbIzO6S9O8kvU3SkKTn3X116P6/lPQpdz9U8XX7JO0Lbg8Pr97wwCf/SpI00idlbHnhqOCu6dL0tJG+4v+n51f2WEkxOzurwcHBuJuRKPRJOfojij6Jok/K0R9R9EnUjh07Trv7xmYfJ/awZmbvk/RuSW9397Ola+ckbQqqa2b2pKTfcvcv1HussbExn5iYqPcpPefYsWPavn173M1IFPqkHP0RRZ9E0Sfl6I8o+iTKzFoS1mKd3V6qjN0g6d8EQa3kLyXdXvqcayT9qKSGV4MCAACkRZxbd2yU9CFJpyQ9asVhxjl3v1bSb0v6hJk9Jykn6X9ZaiUoAABAGsU+DNpKfX19vn79+ribkShzc3MaGBiIuxmJQp+Uoz+i6JMo+qQc/RFFn0SdPn1a7t70pPckrQZt2tq1a/XCCy/E3YxEYQ5BFH1Sjv6Iok+i6JNy9EcUfRJlZivb9LVCqsIa4hGcGBFsT1J5ggTblgAAsHKENSxLZRAL70831JfVn737Kv3GJ58qO0Gi2j51BDgAABpDWMOSgoC21DmnM/N53fqJr1a9XnmyBCc6AAB6wHQrHoSwhohw9WypgFZLrSpboDLAUX0DAKQQYQ2t02j1TKoerNYM9mlqdn7xdhC0Hr5la81h07Ba1TcOuwcA9DrCGpRbKGj3gScaDmi1Kl7rRqJLtrMZi1wPB7h64XBmPq+d9z6uA3u21P2+AACkGWGth+ULrjPTc5GKltTeOWWVAa5a9S3cpuBjKm0AgF5EWOsxSw13xlHFqlZ9O3Lbtki1j0obAKAXEdZ6SL3hzqRVrfpXZRYrblTaAAC9jHe5HpFbKGjnvY9HgtpQX1YH9mzRw7dsTVzoCSpu69cM6sht2zTUly27P6i0vTg1qzPTczozPad8IT3HpwEAIFFZ6wn5gmv3gSfKrnXbUGIjlTZJ2rB2SPfctLkrfiYAABqRrFIKWipfcE3lXKcmpssqakdu26b1awa1bmSgq0LNUpU2STp9dkanJqapsAEAUoPKWkoF89N+8FJerxh/evH6gT1bEjfcuRLhSpskTZ7Lad8jxZ9z3yNPM58NAJAavJOlTL7genFqtur8tA1rh6ruhdatgkrbupEBXTY2og1rhxbvq5zPRqUNANCtqKylSL7guv3Br+n02Zmy6902P20lshnTPTdtjuwbF3zMXDYAQLeispYiZ6bnyoLaUF9Wv70525Xz01Yim7Ga89lOn53Rmem5mFoGAMDKEdZSIBj6DFeU9u+6Ug/fslWrUh7Qqgnms91/8zXav+vKxet7D57Qi1OzKjhDogCA7sEwaJerNvQ51JfVZWMjqa+k1RPMZxsd7tdQX3Zx/t7egyc0M53X5m2zPVFtBAB0PyprXW7yfC4S1A7vvZYQUpLNmA7vvbZsWDRXKIa22x/8GgsPAACJR2Wti+ULrslzFw5A37/ryp6vqFUTDIuemZ7THQ+d1Mul66fPzmjyfC5VK2QBAOlDWOtC+YIvBo/w9hyjq9O72rNZweKDh2/ZqkP/9ZiOni5eD8JumlfKAgC6G2Gty9TanmPD2iGNDvfH1Krukc2YRvou3A420mUnIwROAAAgAElEQVRrDwBAUhHWuky17TnuvvFqJssvw0hfMZyF+zHY2mP9msEYWwYAQBRhrYvkC647Hjq5eJs5aiuTseIGupPnc2XHVO09eEIH9mwh+AIAEoWw1kUmz+cW56ixPUdz6m3tEVQr037qAwCgO7B1RxcIFhSEV34GYQLNqba1x8x8XnsPntDN9x9new8AQOyorCVcrQUFBLXWqdzaI7zClu09AABxI6wlXOWCAomVn+0Q3tqjci4blTUAQJwIawlWbUHB6Op+5lG1UTCXLWzvwRM6cts29a9i1gAAoPN490mofMF1amI6sqCAlYqdESw8COy893G9ODWrM9NzVNoAAB1FZS2Bqs1TY0FBZwULD3be+/jitb0HT0hiA10AQGdRWUugynlqG9YOMcE9Bv2rMjpy27ayCpt0YdEBAACdQGUtYdj4NlmClaIsOgAAxIXKWsKcmZ5j49uEWdxAd/WFFbh7D57Qi1OzhDYAQNsR1hKksqrGPLVkqVx0sPfgCTbNBQC0HWEtQSqPk2KeWrJUO+0gOAAeAIB2IawlFFW1ZArmsO3fdeXiNYZEAQDtRFhLgGpnfxLUkiubMV02NsKQKACgI1gNGrNaZ38i2YIh0d0Hnlgcuj59dkanJqZZFAIAaCkqazHj7M/uVW1IdN8jT1NhAwC0VNsra2b2YUk/K+lVkl7v7t8oXb9c0kFJ6ySdlfQed//WUvelCWd/dr9gSHTD2qHF0B1smssCEQBAK3SisvZXkt4s6V8qrn9U0n3u/lpJfyLpYw3elxqVqz85+7M7ZTOme27aXFZhmzyXo7oGAGiJtoc1d3/M3V8IXzOziyVtlnS4dOmIpFeb2aZ697W7rZ2UL3jZggJWf3a3bMbKNs3d98jTuuG+ryi3UIixVQCANIhrztolkr7n7guS5O4u6XlJly5xXyoEiwqCo4skVn+mwehwvzasHVq8PTOf1+4DT1BhAwA0xYpZqM4nmK0KglNT38hsXNI73f0bZvZGSYfc/YrQ/ccl3SnpXK373P2xisfcJ2lfcHt4eHjD0aNHm21q252dc93z9fzi7YsGTbdckVHGWh/YZmdnNTg42PLH7Wbt7JOCu6ZyKvv3vf31Wa0dSG4Y5zkSRZ9E0Sfl6I8o+iRqx44dp919Y7OP00hY+6GkQ5I+3swk/4qwdrGk5yRd5O4LZmaSvi9pq6Tzte5z9/F632NsbMwnJiZW2sSOyBdcN9z3lcW5au0+qP3YsWPavn17Wx67W3WiT16cmtXegycWbx/YsyWx8xF5jkTRJ1H0STn6I4o+iTKzloS1RoZBt0sakPSYmT1uZr9sZiPNfFN3/6Gkk5J2ly7tlDTu7uP17mvme3ZSsMltteGvaosKkvgGjuasGxlg01wAQEssGdbc/aS73yHplZI+LGmXpBfM7GNmtm2przeze8zsBUkbJX3OzP6pdNetkm41s2clvV/SL4e+rN59iRbMR7v5/uORN2cWFfSOWueInpqYrhnkAQCopuF91tw9J+mTkj5pZq+R9LCkL0rKLvF1t0u6vcr1b0u6rsbX1Lwv6SbP56rut1XtpAKCWroFm+aempheXEwS/H/D2iHdc9NmngMAgCUtazWomb3BzP5M0pcl5STtbUurUigc4iROKegV4U1zw4IgDwDAUpasrJnZqKQbVRyKfKWKiw3+dan6hRVo96ICJEuwae7k+Zwmz+UWq2sMhQIAGtFIZe27Ki4y+F8lbXT33yKoNWd0NcdJ9ZpsxrRuZKBs49w7HjpJYAMALKmROWt/IeklST8l6aesYi8wd/+tNrQrVSoXFqB3jQ73a6gvq5n5vGbm8zozPaf1a9iXCABQWyOVtZslvVXSjIob1lb+hzqqnVaA3pXNmO6+8erF23sPntCLU7NU2BCbYKuh4L/guVh5vbDEnpwA2qeRytrbVQxsN0p6RMXNcf+5ra1KERYWoFKwB1uw397egydYHYq2yxc8sqglX3Dd8dDJxeeiVNz/8c/efZV+45NPlV2fmc5r87bZms/R0WGmdwDtsmRYc/fPS/q8ma2RdIOkB81sRtL73f2JdjcwTVhYAOnCHmy7Dzyx+GZ4+uwMQ6JoiUZDWS0z83nd+omvRq7nCio7laPSUF+27t6RhDlg5Zazz9qUmf03Sf9K0nsl/YQkwtoysLAAgWp7sN3x0Ek9fMtWniNYsdxCoeyPgEbVqqaFr7+8xGPMzOcbDnMEN2B5Gtm6IyvpZ1XcuuPVkj4habO7f7/NbUuFl8/Px90EJFSwB1t4wcGpiWmNru7nzQx11aqe1QtLUu3qV/B8e/iWrWWPG75+9NPHdP3111RtSyNVu3CYq9YOnvNAbY1U1k5Lel7SxyU9Vro2Wtp/Tc0c7t4L/vDT/xB3E5BgwYKD4E2MEw5QSxDQGg1HB/ZsqRnKagm2mKl2fU1/9fskRUJeZbsr21utCkflDaitkbA2K2lM0m9L+i1J4d8gl3RZG9qVOiwsQC3rRga0Ye1Q2UKU8FFl6G3BqsxG55wN9WV1eO+16l+1rANqmlIr5AWCMFcvaNaqvBHcgMYWGGzqQDtS6U9/4Q2LL2C84KAWTjhANUuFtKWGNJMkHOYqq3BLVd4IbsAyFhhg+daNDFAZQUOqVSb2HjyhA3u2aN3IAG9QPabWQoE0BJdqz/V6lbd6c926tQ+A5SKsAQkSPuFAYg+2XhFeMFBtoUAQUtIa3KtV3hqZ6xbHkC8QB8IakCC19mBj/lo6NTIfrdeqq40GN6kY3nbe+3jZYgqqbUgjwhqQMNX2YJs8l+NNKGWCo+jCC0vCqBrVnutWWX2srLZ1+1Ax4le5Pc6awT5Nzca3FRdhDUigbMY0uvrC6uF9jzzNcGiK5AuuUxPTZUGN+Vj1Vc51O3Lbtqrz+irnuB3ee62yGdPkef7gQVG1fQor72909XWnENaAhBod7i/b0uP02RmdmpjmyLIuVW+fNI6iW76gAh2utlWb47bz3sfVvyqj3EKBamUPqRXIkhjEGkFYAxIq2NIjPBxKha071TsGasPaIYLaClVW28Jz3MJDo7mFgqTqc9wCVN26R6cqY5XHsC11/m01Y+9tqgmLCGtAggVHUlVW2Dj0vTsECwiqHQOV9hWecQiHt1rDpFL1A+mpuiVTZTBrZRBbKnhVHsMWZ6AnrAEJV63CFuzBxiTq5Kq1gIB/t86oHCZd3b9Kez7+ZM03+cqqG/8+8ajcxqaZYFYvkC3n33epEzo6gbAGdIHKQ9+lC9UBhkWTI/xGM3kuF1lAQOWms2oNk4bVWlnKgo/2acV8suVUxtKAsAZ0iWp7sEnsw5YU9ealsYAgGWpVSKoNmVbbgDccDgrOcXCNarZaVi2YpSmINYKwBnSR8NAO54gmR77gLCDoYuHfq3ob8IbD28x0Xpu3zVJ5K8kXXFO54hzNyuvNhrNe71uJsAZ0nWrVgTseOqmHb9na8y9onRZUDCbP5crejNhRv/vU24C3WtjIFWpvxhtI6799rUn/P3gpr0Pjxxt6jFbNJ+sVhDWgS4XPEZ2Zz7MHW4fU2y9NKgY1Vup2t3pz3YJ/95crvqay8iZVDyThnfCTGEratS0G1bLmENaALpXNmO6+8erFN4h9jzzNdhBtttQRURvWDjF3MIWqhbejnz6m66+/pm54qRbgwmpV4yS1ZauIdu9P1p9R1T3sJMJZswhrQBdbNzJQtgdb8ObACtHWCg93VgY1zqLsPdmMaU1/9WFTqfHQUy3MBauFgxMXlrsJay2t3rm/WtuOf+lRqsptQlgDuliwB9uZ6bmyF2JWiLZOrWra/l1XanR1PwENVeeR1gtw4UBWKXxtqcpcu6x0W4yM8XvQLoQ1oMtlM6b1awb18C1byzbOnTzHwdXNqnbgusQKTyytXoALD3UGOnVmZa/tT5YWhDUgJbIZ0+jq/sXbwRw2NmJdvoK7XpyarXrgOtU0rFRlgKtXjQsvRGglnrvdibAGpMjocH9kDhtH6CxPbqGgD53Ma+i58uEnqmlot6XCHHoXYQ1IkfActmpH6LDwoL5gc9tcQRoqXWOFLYC4EdaAlAnmsFU7QoeFB1GV53lWbm5LSAMQN8IakFIcTbW0evumsbktgKRg1jGQYsEcmPDCg70HT+jFqdmeDm35QvEMw2orPSXposHqB34DQByorAE9IHw0lVQMbMFK0V6TWyhUPXQ9WOkpFTf3ZOgTQFJQWQN6QDZjOrz3Wg31ZRevBStFF3qgwhZU0l6cmtXOex+PBLVgpee6kQGtGxlgc08AiUJlDegRwRy2ypWi//lreW35qdnUbu1Rq5ImiS1NAHQFwhrQQ8IrRXfe+/ji9SC8pWUT3WCFZ77gVY/rScvPCaA3ENaAHtS/KrO4tcfLoevhTXS7ccuKYLiz1rE9VNIAdCPCGtCjgmHRo58+pre8dUtkE91uqz7V24aj234WAAhL7CuXmV1uZo+b2bNm9qSZvS7uNgFpk82Y1vRfGBqttgAhydt8BJW0attwDPVldWDPFt1/8zV6+JatBDUAXSvJlbWPSrrP3R8ws3dJ+pik62JuE5BatRYgBFW2pB25VG/hwP5dV3KOJ4DUSGRYM7OLJW2W9I7SpSOSPmJmm9x9vNbXFbz4V3acJs/lYv3+QDNqHVU1M58vC21xzPsKHwtVa+GAxIHrANInkWFN0iWSvufuC5Lk7m5mz0u6VNJ4rS+anpduvv94Z1oIpFi4yhaerB+ENkllwU1SW8JbeFVnrUUD0oWFA+1qBwDEydyTNxfFzN4o6ZC7XxG6dlzSne7+WOjaPkn7gturRkY3XPf+T3S0rbVcNGi65YpM7Jtrzs7OanCQ8w3D6JNyS/VHwV1TOenPv5lXrlD7cfoz0q9ckVWQk0b61NDzv+Cu6flq1xv7nv/hqqxWtTic8RyJok/K0R9R9EnUjh07Trv7xmYfJ6lh7WJJz0m6yN0XzMwkfV/S1nrDoBetW+ffHj/doVbWl5S/7o8dO6bt27fH3YxEoU/KNdofjVa5ApWVt1qP2chj1XrMdv2e8RyJok/K0R9R9EmUmbUkrCVyGNTdf2hmJyXtlvSApJ2SxusFNan4VzyHLwPtERwKL0kP37K1bP5YtcAVHjJtRpzz5AAgCRIZ1kpulfSAmd0laUrSnpjbA6AkHNykxsJbPfWqcAQ0AL0usWHN3b8ttuoAukK98NYIAhkA1JbYsAage1WGNwDAyrGlNwAAQIIR1gAAABKs7cOgZvZhST8r6VWSXu/u3yhdv1zSQUnrJJ2V9B53/9ZS99Vz9uxZbdzY9ArZVJmbm9PAAMNRYfRJOfojij6Jok/K0R9R9ElVG1rxIG3fZ83M3iLplKQvSXpnKKx9XsWNb4OzP+909+uWuq+esbExn5iYaNvP0o3Y9yaKPilHf0TRJ1H0STn6I4o+iTKzvLs3XRhre2UtOHHAQjuZ1zv7U9L5Wvcttc9aNwmfc9hOU7n4z0tNmk71CSscAQCtENdq0Hpnf56rc994TO1tqXzBdfuDX9PpszNt/14vT+V1aJzzUsM61SdDfVkd3nut+lcxNRQAsHIdO27KzMZVGgatd/animFtyXNBS9fLzgYdHh7ecPTo0fb+IC0wlXPd/UzjG4Y2wwsuo7pTppN90p+R7rw6G/sZsfVwnl8UfRJFn5SjP6Lok6gdO3a87O5rm32cuCpr35W00cxWhc7+vETS8yoOg9a6r4y775e0P7g9Njbm3TBefmZ6brGys3/XlRpd3d+27/Xoo4/q+uuvb9vjd6N290m+4GXHLF3z5msSvecY80yi6JMo+qQc/RFFn1Q13YoHiSWsLXX250rOBe0m+cKFaubo6v62vpGv6Wdz0kqd6JMDe7YsBrbwvzcAAMvV9sk0ZnaPmb0gaaOkz5nZP5XuulXSrWb2rKT3S/rl0JfVu6+rBecmIt3CCwvueOgkgQ0AsGKdWA16u6Tbq1yvefZnms8FnTyfWzzgeqgvq9Hh9g2BIj6jw/0a6stqZj6vmfm8Tk1M67KxEVaHAgCWjWVqbZAvFLeGWKqacveNV/PmnVLZjOnuG69evL3vkad1+4Nfo8IGAFg2DnJvsfC2HBvWDumemzbXDGQEtXRbNzKgDWuHFrdoOX12RpPnc8whBAAsC5W1Fps8n4u8OQfyBdfkufZvhItkyGZM99y0Wft3XRl3UwAAXYzKWod0ciNcJEc2Y23dmgUAkH6EtQ4JV9wkacPaIRYX9KCgsspRVACARhHWYrB/15WsDOxR+x55WpKWnM8IAECAOWsxGF1NVaWXjA73a8PaobJrlfMZAQCohcoa0GbBQoPJ8zlNnsstVtcAAGgEYQ3ogGyGY78AACvDMCgQEzbIBQA0grAGxIQzQwEAjSCsAR0UnBkqSTPzeRYZAACWRFjrAE4uQKDyzFAqawCApRDW2iw4uYAVgAiEt21hKBQAsBTCWptxcgEqMRSKbpAvuM5Mz/HHBJAAbN3RQZxcAOnCUOjegyckFY+g4vgpxCVf8MgfDPmC646HTmpmPq+hvqzuvvFqTeWK4S34Y3PyPM9boFMIax3EyQUIhJ8H+x55muOn0HFB5SwIZbXMzOe19+AJvTyV16Hx4+pfVRyQyS0UFoNcNmMEN6CNCGtADIIjqIIh8uD4KTbORastVTlbrtxCYfHjIMhJKgtukghvQAsR1oAYBEdQnZqYZvEJWi4IaI2GssqgFVgz2Kep2XlJ0t9+/lEd+V625mOFg1u1xyS8AStHWANiks2YRldfWGzCRG40q9GhzUAQqNaNDNQMUkG1d+2A6eFbti5W6YIgVysQVoY3hvqBlSOsAQlxx0Mn9fAtW3kzw4oE2wSFV58HalXOllvtqjzjNvg4HOJqhbfTZ2d0amJao6v7qbIBy0RYA2IUbOMxM5/XzHyeNzMsS3g+2uS58m2COjn5vzLEhcPb5Lnc4lB/8P+hvqwO7712cbECgPoIa2328vn5uJuABKvcxiN4M2PICPUsNdwZ9zZB4fBWuZhGKg6R7rz3cR3Ys4U5bUADCGtt9oef/oe4m4CEWzcyEHkzY3Uoaqk33CkVg36S9nMMFtMECx7C89iY0wY0hrDWIZxcgFrCb2bhISMgLBjyrDfcKSWzQhWutB25bZt2H3iCOW3AMhDW2uhPf+ENZUMBvPiglso5P0BYrWpa3MOdK9G/KsOcNmCZ+E1oo3UjA4v/ddOLKYDkyBdcpyamI0EtacOdyxH8cbJuZECXjY1ow9qhsvuDOW0vTs3qzPTc4n9sb4NeRWUNSCjemFCtorZ/15WpGipsdE6bxLw29C4qa0BC3fHQSQJbjwpWe1ZW1IJqWtqq9UGlbf2aQR25bZuG+rJVPy+Y18bvBXoNlTUgQSr3XWNFaO9J0/y0laic0yZF57VRYUOvobIGJEiw71pg8lyOKkKPqFVNk7p7ftpKhOe0VZvXRoUNvYbKGpAw4Tdkqgi9oV41LU3z01YqmNd2amK6rMLGqlH0Cp7hQMIEO74Hgg1ykV5npudqVtPSNj9tpbIZi1TYKleNUmlDWlFZAxKmWhUB6RQMfYZXPVJNqy343ajss+BjKm1IK57RQAJlM6bR1RdOvGDuWvoEQ5/h0DHUl6WatoRsxmquGp2Zz2v3gSf4XUGsgj/CzkzPtewxqawBXYC5a+kzeT56bNThvdfy79ug8KrR8P5sM/N5jq1CywTHvK0Z7NPU7HzZfcERkuFpKvmC646HTkaOU2sWYQ1IqGDuWvCGzuHu6ZEvuCbPXXiB75VtOVotfEzbgT1bFgMbx1ZhOWoFsqWCV/C8yi0U2t5GwhqQUMxdS6dqKz9HV1MBata6kYGyP26kC8Oih/deq6nZeSptkHQhnAUfr7QS1omQFmgqrJnZQXff06rGAChXOXcN3a9y+HPD2qHF4RSsXK1jq4IVoxLHVfWaIJSFhyubCWdDfdnFfTDDjxFcDz+vgird2Htb8IOo+cra9c18sZmNS5ot/SdJf+zuf2Fml0s6KGmdpLOS3uPu32rmewFpMHkuR3WgSwVvHAx/tk+tYdHA6bMzOjM9p/VrBuNoHjoot1DQ7gNPaGY+v6zhynrBK/zaGz5lo9ZrciunrCwZ1szsh7XukrS2BW14l7t/o+LaRyXd5+4PmNm7JH1M0nUt+F5AV2OhQXeqtektw5/tU21YVCpu83Fgz5Zi1Zo/fFKjcmgzHNSrhbTKUFYtkIVVBq/wHwad0EhlzSS9TdLLVa7/fasbZGYXS9os6R2lS0ckfcTMNrn7eKu/H5B0LDTofrU2vWX4s30qh0V/7cGvLb5ph/dlC96wCW7dIxzMgtuNDG0u9e+d5NfURsLaVyVd5O7PVN5hZj9oQRseNLOMpCck/Y6kSyR9z90XJMnd3cyel3SppPEWfD+gq1RbaMA+Ut2BTW/jFa5+PPwrWxeHxQIz8/nFfxsq1skX/D41OudsqC+rg//+TTqXW5BUe7iyG5h7/Rd9M1stKefu82Y2JmnG3adb8s3NLnX3582sT9IfSnq9pN+VdMjdrwh93nFJd7r7YxVfv0/SvuD28PDwhqNHj7aiaSs2lXPd/UzxSXTHG7Ja0x/vE2N2dlaDg8zPCOvWPgk/t/oz0p1XZ5Wx5p9f3dof7dSKPim4675vFvTS7IXX2Fb+u3VaGp4nBXdNz0sFl/78m3nlKkbHbn99VmsHGvu3SUN/tFqr+yT49yp+XP3fLKw/I/3KFVkFeWykT7H/ru3YseO0u29s9nGWDGuSZGa/JukDkn5Ukkv6pqR97v45M1vr7mebbojZj0l6VtJrJD2nYjVvwcxM0vclbV1qGHRsbMwnJiaabUpTzkzP6eb7j0uS7r/5mtjLqseOHdP27dtjbUPSdGuf5AuuG+77yuJflK16fnVrf7RTK/ok/Fogdf+eX2l7noQXfIS3xml0Plva+qMVWtUnjVTQqi0ESGLlzMxaEtYaWWDwK5J+XdIvS/py6fI2SR8ys9+U9MeS3rjcb1yq2PWFgt4Nkk66+w/N7KSk3ZIekLRT0jjz1dDrshnT3TdevThsE6wqTOILVC9j1Wd3CIZIR4f7NdSXXQwFnDMaj+D3Zqn5Z0FI67Uj2RqZs/ZeSTvc/fnQtc+Y2bdUrITtX+H3Xi/piJllVVyscErSL5Xuu1XSA2Z2l6QpSezlBkhlL05BNYC5NsnBqs/uk82YDu+9tup8tp33Pq4De7b0XDDohEY3pq2soPXqH6eNhLVMRVCTJLn7uJmNu/v7V/KN3f2UpKtr3PdtsVUHEFG5MlRidWiSsOqzO9U6Z1QqVtqosrVOo4sEerWCVksjYa3fzAbdfTZ80cyGJPHMBToovB1B5VwbxCu3UGDVZxcLrxw9ctu2skpbuMoW/FsWGpjvjaLlDnGylUpUI2HtqKRPmNmvBPPLzGxU0n0q7oEGoIM6vRkjlpYvuHYfeGLx9lBfljlqXSyotFVuuxL+eFWuoHe8w/k3rqGRgMYQZ+MaCWsfUPHopxfM7DkVV4O+VtKnSvcBiBn7rsVr8nyu7M3o8N5redPpctmMaf2awUiVLfDSrOvUxDTV05J8wTWVKw5xskig9ZYMa+4+L+lGM3uNiicLSMVVm//U1pYBaNgdD53Uw7ds5YUvAQ7s2cLcphQJz2eTVDb9IPh/r1aIKqtnP3gpr0Pjx6t+LkOczWn4IHd3/2dJ/9zGtgBYhvCWAzPzeRYZxKDaNh28CaVPeOpBsMjn5akLJzCGT0KQ0l85Wu4iAQJa8xoOawCSpdq+a7wgdk5uoVB1eAzpFizyOfrpY3rLW7dUDSxBeEvDcUfLOYezP6OyRRjd+PMmFWEN6GKV+66x51r7VTvvM8A2Hb0hmzGt6S/OaQsPkVYGmZn5vHZ99MuLX9cN4a3R/c8C4erZ8S89qvVrOIKrHQhrQBer3HeNPdfaq9amt40eUYT0qVydXW0VaaBaeAsWo0yej6cyvtxwFqg21Bv3OZxpRlgDulgwJHNqYnpxsjPDoe0zeT5XFtTYLBWVqq0i7V+VUW4hegJ5sH9bcP9Kw1sQuJb6/OUMaYZ1yzmcaUZYA7pcNmMaXX1h6I3h0PYJb5HCeZ+oJ7yKdHS4f3EvvmrhLfi4WnirDEmVwoGr3uevpGrG3LPkIKwBKcBwaHtVm6fGeZ9YSniINJuxmuGtUji8VRtOrWW5nx8gnCUfYQ1IgWrDoWiNavPUhvqyLCTAstUKb5K0un+V9nz8ybavLmZIszsR1oCUqBwORWvUmqfGmxuaVW1xQhDe1gz2aWp2vuHHavTzCWbdibAGpBQLDZqXL3jZhrfMU0M7VYa35U5jYNpDehHWgJRioUFzqg1/Mk8NQBxYbw6kSLDQIHD67IzOTM/F2KLuVPDiId3hoMaGtwDiQmUNSJFqCw045H158gXXfd8saOG5Cws1GP4EECcqa0DKZDOmy8ZGNNSXlVRczn9qYlpnpufK9glDdWem5/TS7IV+2rB2iKAGIFZU1oAUqjzkPaiyMYetvmDj0AAVNQBJQGUNSKl1IwNl89ekC5vlorrJ87nFfa6G+rIENQCJQGUNSKlg/trk+Zwmz+XYLLeO4MzE8DYdSx3zAwCdQlgDUqxy3yZEVduiQxJBDUBiMAwK9BgWGZQ7Mz0XCWoXDRrbdABIDMIa0GPueOgkga2k2oKC+2++RrdckaGyBiAxCGtADxgd7o9s5dHLgS1fcJ2ZntOpienIgoJ1IwPKGEENQHIwZw3oAdW28gi28eg1teaosaAAQFJRWQN6ROVWHr16FNXk+VwkqG1YO8RCDACJRWUN6BG1jqK6eVNvDYeGh3/377pSo6v7NTrMAe0AkovKGtBDqh1FNdVDe+RWLigYXd2vdSMDBDUAiUZYA3pMMH8tcM/X83pxajbVZ4fWWlDA9hwAugHDoEAPWjcyoKG+7GJwCRYepPHs0NxCQbsPPLH4swZYUACgW1BZA3pQNs6KhWYAAAvtSURBVGM6vPfaxeHQQNrODs0XvGpQY0EBgG5CZQ3oUf2rMnr4lq06+uljuupNV6bu7NB8wcuGPSXpwJ4tymaMBQUAugphDehh2YxpTb9pdPWFuVtpmLdWbejzwJ4tWr9mMMZWAcDKMAwKoMzegyf04tRs14a2akOfDHsC6GZU1gAsHkcVXnDQjYsNag19sj0HgG5GZQ1A1QUH3XbCQXCMVHjuXTD0SVAD0M0IawAkXVhwsH/XlYvXumVINKiohY+RYugTQFowDApgUfiEg24YEg02u73joZNlQ5/7d12py8ZGEtdeAFiJxFbWzOxyM3vczJ41syfN7HVxtwnoBbWGRE9NTCeqwpZbKOiG+76ivQdPRBYTENQApEmSK2sflXSfuz9gZu+S9DFJ18XcJqAnBEOi4UPf9z3ydGIqbNVWfA71ZXX3jVezmABA6iQyrJnZxZI2S3pH6dIRSR8xs03uPl7r6wrusU+InjyXnt3f0duCIdENa4cW54IFFbbwvmxxmDyXY8UngJ6RyLAm6RJJ33P3BUlydzez5yVdKmm81hdNz0s333+8My0EekA2Y7rnps2RCluSsNktgLQz9+TMQQmY2RslHXL3K0LXjku6090fC13bJ2lfcHvVyOiG697/iY62tZaLBk23XJFRxuL9S392dlaDg7yRhdEn5Rrpj4K77vtmQS/NJuv1ol2/ZzxHouiTcvRHFH0StWPHjtPuvrHZx0lqWLtY0nOSLnL3BTMzSd+XtLXeMOhF69b5t8dPd6iV9SXl7MFjx45p+/btcTcjUeiTco32R77giTvkvV2/ZzxHouiTcvRHFH0SZWYtCWuJHAZ19x+a2UlJuyU9IGmnpPF6QU2SMmbsqwS0STbD7xcAxCGRYa3kVkkPmNldkqYk7Ym5PQAAAB2X2LDm7t8WW3UAAIAel8g5aytlZguSfhB3OxJmRNJ03I1IGPqkHP0RRZ9E0Sfl6I8o+iTqR9296cJYYitrK/SDVkzkSxMze4E+KUeflKM/ouiTKPqkHP0RRZ9EmdkLrXicxB43BQAAAMIaAABAoqUtrO2PuwEJRJ9E0Sfl6I8o+iSKPilHf0TRJ1Et6ZNULTAAAABIm7RV1gAAAFKFsAYAAJBgqQhrZna5mT1uZs+a2ZNm9rq429QJZvZhMxs3Mzeznwxdr9kfae4rMxs0s/9a+tmeMrP/x8w2le67uHT7OTP7hpm9OfR1Ne9LAzP7rJk9U+qTL5rZVaXrPfk8CZjZ74V/d3q5P0qvI/9Yeo48ZWa/WLrey30yYGYfKb0ufNPMDpeu91yfmNna0HPjqdLPt2Bm/6rHX1u3m9lXzexk6efbU7re+j5x967/T9LnJb2n9PG7JH057jZ16Od+i6SNksYl/WQj/ZHmvpI0KOlndGEu5q9L+mzp449L+v3Sx9dI+hdJq5a6Lw3/SVob+vjnJH2tl58npZ9ps6S/Kf1b/yT9Uf4a0sjP3QN98n9I+nDo9eTHer1PQj/n+yR9qvRxT762SjJJL0l6Q+n2Jkmzkn6kHX0S+w/cgg67WNLZUEeYiqcYbIq7bR3sg8UX2nr90Wt9JWmLpH8qfTwtaSx035OSfnqp+9L2n4pn7J7o5eeJpAFJX5b06uB3p5f7o/QzLb6GhK71bJ9IWl36+Ubok6r9801JP1f6uCdfW3UhrL2ldPsNkk5L6m9Hn6ThBINLJH3P3Rckyd3dzJ6XdKmKL0C9pl5/nKtz33hM7W2n90r6lJldJCnj7hOh+8YlXVrvvo61sgPM7JCk60s3d6i3nyd/IOmwu3/HzIJrvdwfgQfNLCPpCUm/o97uk9eo+Eb8ATN7u6QZSb+vYhjr1T6RJJnZdZIukvTXvfzaWvr33SXpqJmdkzQq6X9WsbLW8j5JxZw1SZX7j1jVz+od9fqjJ/rKzO6SdLmk/1i61NN94u6/5O6XSPqApD8NLld8Wur7pPRGc42k/1Ll7p7rj5C3uPuVKg4PvyTpYOl6r/ZJn6TLJH3L3beoOKXikyoe0dirfRL495IOBaFUPdofZrZKxT9q/id3f5Wkt6mNvzdpCGvflbSx1HGy4p/Kl0h6PtZWxadef/REX5nZ+1T8C+ffuvt5d3+pdH0s9GmvkvR8vfs61d5OcveDKlbYXlBvPk/eKuknJH3HzMZVnPN5TMWh0F7sD0mSuz9f+v+8pD+T9K/V268l/yKpIOlBSXL3pyV9R8XXhl7tE5nZakm/qOK8K/X4a+tVkl7p7n8vSe5+XNL3VBwObXmfdH1Yc/cfSjopaXfp0k5J4+4+HlujYlSvP3qhr8xsn6QbJP0bdz8buusvJd1e+pxrJP2opC81cF9XM7M1ZvbK0O2fV7Fy0pPPE3f/oLu/0t03ufsmFUPr9lKI7bn+kIpvwGa2NnTpBkkne/m1xN3PSPpbSdslycxepeIcxy+qR/uk5BckPePu/xi61pOvrboQzn9ckszsv1Nx+PxZtaNPOjkhr13/SfpxFScMP6vi5Okr4m5Th37ue1R8s1lQcSJrMJm+Zn+kua9UrJK4pH+W9FTpvydK962X9FlJz6k4Ofatoa+reV+3/6fiX/ZPSvq6pKclfU7SVb38PKnon3FdWJzTk/2h4nDfSUnPlJ4n/7dKk+J7tU9C/fKFUp88Jenn6RN9UdLNFdd68rW19PPdEHptfUbSu9vVJxw3BQAAkGBdPwwKAACQZoQ1AACABCOsAQAAJBhhDQAAIMEIawAAAAlGWAPQEWb2VOm/b5nZQuj2X5jZFjN7sANt+AUzu7fK9deb2efN7Gkz+4aZHTezn2x3e5ZiZl8ws3fG3Q4A8UrD2aAAuoC7XyVJZrZJ0ongdshNHWjGz0k6VOX6Q5L+o7v/N0kys0skzXWgPQCwJCprAGJnZj9tZidKH28yszNm9odmdtLM/rFUebvPzJ4xsycrTmV4X+na18zsM6WgVe179En6KUmPVrn7UhU3mJYkuft3vbgjvczsR8zsz0vf4xkz+z9LjyUz22Bmf1W6/oyZ/afS9fVm9n+Z2ddLlbpbQu0YN7PfM7PHzew7ZvaB0H2vM7MnSj/Lg5IGQ/d9wMz+IVSRfNWKOhtA1yGsAUiiiyR92d2vlvQxFU9e+C/u/gYVd4b/dUkysxslvVbSde6+WdLDkj5S4zGvl/S4u+eq3PcHkh4zs781sz8ys6tD931I0mPu/iZJV6o4IvHrpfsOq3hKxhtKbftw6fqHJf2ju79e0v8g6XfN7E2hx1zr7tskvUnSb5rZhtL1T5R+zs2S7lbx0HmZ2aik90naXKpIbpP0Ys3e+//bu2PXKIIwDOPPh1oIEVEQWwVbU4mkEkWDbWJnoY1gISKKIHYWaum/IEgCkiaKkEaCCrGIRcIlJlgZEGw0XUwhYvgsZk6P43Kx3BzPr5rdmXx724SXmVlG0kBxGVRSE21m5kxtLwJfM7NVrxeA0doeA04BC+XcbPYAW9vUHANe9OrIzCcRMUkJVmeAuYi4lplT9e9GIuJuHb4f+BURQ5TQNNpRZ702L1CCHZn5PSKmgfOUo7/g3wHh6xGxBhyPiB+UA+Unat98RHys4zcox9NMRsRrYCYz/84EShpshjVJTdS5X2wL+Nl13f7fFcCjzHzar1iUJHcRuLfdmMz8RpmZex4RXyh76KbqM8Yyc62r5tAO79B9ll/n9Xbv0/P8v8zciogRSjg8C8xHxOXMnNvhN0gaAC6DStrNXgE3IuIwlH1pXUuYbaeBT5m52atIRIx37EPbCwwDnzuecb/eJyIORcSJWus9cKejzpHanAWud9wbB970e5HM3ABWqB9a1GXTk7V9ADiamXOZ+bA+t9d7ShpAhjVJu1ZmTlD2jb2LiCWgRdmb1m0ceNmn1CVgJSKWgSXKzN6D2ncb+A20av8scKz2XaEska7W57f3st0Chuv4t8DjzGwvgfZzFbgZEYuUsPeh3j8ITNcPFpaBfcCz/6gnaQBEZs9Zd0kaGBGxCpxrf+EpSbuJYU2SJKnBXAaVJElqMMOaJElSgxnWJEmSGsywJkmS1GCGNUmSpAYzrEmSJDWYYU2SJKnBDGuSJEkN9geOnUK+j95eWQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1122626d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TCLab Model disconnected successfully.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1122626d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from tclab import clock, setup, Historian, Plotter\n",
    "\n",
    "TCLab = setup(connected=False, speedup=10)\n",
    "\n",
    "controller = PID(2, 0.1, 2)        # create pid control\n",
    "controller.send(None)              # initialize\n",
    "\n",
    "tfinal = 800\n",
    "\n",
    "with TCLab() as lab:\n",
    "    h = Historian([('SP', lambda: SP), ('T1', lambda: lab.T1), ('MV', lambda: MV), ('Q1', lab.Q1)])\n",
    "    p = Plotter(h, tfinal)\n",
    "    T1 = lab.T1\n",
    "    for t in clock(tfinal, 2):\n",
    "        SP = T1 if t < 50 else 50           # get setpoint\n",
    "        PV = lab.T1                         # get measurement\n",
    "        MV = controller.send([t, PV, SP])   # compute manipulated variable\n",
    "        lab.U1 = MV                         # apply \n",
    "        p.update(t)                         # update information display"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.4 Analysis](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.4-Analysis)",
     "section": "4.1.4 Analysis"
    }
   },
   "source": [
    "## 4.1.4 Analysis\n",
    "\n",
    "The good news is that the controller did steer the heater to the desired setpoint. There was some overshoot and undershoot along the way, but eventually the system did get there.\n",
    "\n",
    "In examining the response more closely, however, we can see at least two issues that must be addressed before this control algorithm could be applied to large scale indusrial control:\n",
    "\n",
    "1. A sudden adn abrubt \"kick\" to the manipulated variable when the setpoint changes.\n",
    "2. A mismatch between computed and actually feasible control actions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.4 Analysis](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.4-Analysis)",
     "section": "4.1.4 Analysis"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "nbpages": {
     "level": 2,
     "link": "[4.1.4 Analysis](https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.html#4.1.4-Analysis)",
     "section": "4.1.4 Analysis"
    }
   },
   "source": [
    "<!--NAVIGATION-->\n",
    "< [4.0 PID Control](https://jckantor.github.io/CBE30338/04.00-PID_Control.html) | [Contents](toc.html) | [Tag Index](tag_index.html) | [4.2 PID Control with Setpoint Weighting](https://jckantor.github.io/CBE30338/04.02-PID_Control_with_Setpoint_Weighting.html) ><p><a href=\"https://colab.research.google.com/github/jckantor/CBE30338/blob/master/docs/04.01-Implementing_PID_Control_with_Python_Yield_Statement.ipynb\"> <img align=\"left\" src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open in Colab\" title=\"Open in Google Colaboratory\"></a><p><a href=\"https://jckantor.github.io/CBE30338/04.01-Implementing_PID_Control_with_Python_Yield_Statement.ipynb\"> <img align=\"left\" src=\"https://img.shields.io/badge/Github-Download-blue.svg\" alt=\"Download\" title=\"Download Notebook\"></a>"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}