Learning Center
Plans & pricing Sign in
Sign Out



         Project Proposal

Storm Restoration Predictive Model

           Tamer Rousan
         Mike Pingsterhaus
            Scott Pecucci
        The primary mission of power companies is to provide a reliable, economical and safe
supply of electricity to customers. Severe weather conditions such as lightening storms, ice
storms, hurricanes, etc. have the potential to wreak havoc with the electric utility transmission
and distribution systems, causing massive power outages, taking from days to weeks to repair.
This in turn can affect business operations, heating, water distribution, traffic signaling and
countless other aspects of daily life. It is estimated that an average storm can cost an average
utility (Ameren) about $1,000,000 including material and labor.
       To the extent that a company knows ahead of time how long outages will last, it can better
inform its customer, the public and the state utility commission. Knowing the restoration time
with some degree of certainty allows people to plan appropriately, therefore, providing better and
earlier information can allow customers to be prepared. The objective of this project is to build a
storm model that can be used to predict outages due to ongoing or approaching storms. An
accurate storm model holds the potential for lowering operational cost and reducing customer
down time.
       In addition to the predictive model, a weather station with data acquisition capabilities will
be designed. Providing instantaneous measures of the weather conditions and serving as an input
to the model. The weather station should be capable of providing wind speed, wind direction,
humidity and temperature.

Benefits for Power Utilities:

      Reduce customer down time

      Lower operational cost

      Better inform customers and the public of expected post-storm restoration time.

      Plan more effectively the management of crews and power restoration.


      A comprehensive Storm outages database

      Predict the number of outages that are going to occur as a function of time for a given
       storm path

      Predict the most likely damages for a given storm path

      Data acquisition of weather conditions for a given storm path

                                    DB2 Tables

                                      SAS                            ASOS
                                   Datasets and                    Weather Data


                                   Visual Basic

                         Figure 1: Data-base Chart Flow

OAS DB2 Tables: OAS (Outage Analysis System) is Ameren’s central outage database in which
daily outage data is stored. This database contains detailed information regarding every outage,
information such as outage type, count, duration, cause, etc.
SAS Datasets: SAS (Statistical Analysis Software) is an integrated system of software products
that allows a programmer to perform data mining, retrieval, management and statistical analysis.
SAS code will access OAS DB2 mainframe tables, and retrieve needed storm data (we are
interested in the outage type, outage cause, outage component, outage duration, outage area and
outage counts). SAS program will then be developed to create SAS datasets that will query,
analyze, sort and provide some statistics (mean, average, percentage, standard deviation). Storms
from 2006 up to date will be our main focus. Output will be entered into a databse.
ASOS Weather Data: ASOS is a network of 12 weather stations scattered across Illinois. 1 and
5 minute surface weather observations are collected. Surface observations of temperature, wind
speed, pressure and precipitation are reported. Storm weather data from 2006 up to date will be
our focus. Data of interest will be then entered into a database.
Visual Basic GUI: Visual basic 6.0 will be used to develop a GUI (graphical user interface) with
the database.
Access Database: The output of the SAS outage data reports and the ASOS weather data will be
entered and maintained in Access created a database. This database will provide graphing
capabilities to help with data analysis and pattern observation.


                                AC Voltage 120V 60Hz
                               DC Voltage 5V output at 1 A

        Temp                                                        RTC

                                          PIC                   CMOD
        Humidity                                                                Ameren’s
                                                                or DNP            RTU
        Sensor                      Microcontroller            converter         RS232
        Wind Speed
        Wind Direction

                            AC Voltage 120V 60Hz
                              DC Voltage 30V

                                  Figure 2: Block Diagram
AC – DC Converter: The power source will be 2 AC adapters, it will convert 120 VAC to 5
VDC and 30 VDC to power up the circuit component.
Temp Sensor: precision temperature sensors LM135, operating for any type of temperature
sensing between -55°C to 150°C range, with typically 1°C error over 100°C temperature range.
Unlike other sensors LM135 have linear voltage output.
Humidity Sensor: 2bhih-3610 laser trimmed thermostat polymer capacitive sensing element,
with accuracy of +/- 2% error over a range of 0-100%. Has linear voltage output.
Anemometer: Campbell Scientific product (Ameren might donate), capable of sensing both
wind speed and wind direction. Can withstand up to 280 Km/hr with high accuracy up to +/- 5%
error. 24~30 VDC input.
PIC Microcontroller: This is the central block which monitors and controls the work of all
other blocks, interfacing with sensors and talking to Ameren’s RTU or modem (RS232). This
PIC has to have DNP or MOD protocol outputs in order to be able to connect to Ameren’s
Utilinet network and transfer the information through its RTUs. A converter chip will be used.
DNP converter chip: Usually a small chip that can convert the analog output of the PIC to DNP
protocol allowing it to interface with the rs232.
Ameren’s RTU (RS232): Communication Module that lives inside a capacitor bank on
Ameren’s power lines. The RTU uses Utilinet network to communicate with Ameren’s central
database. It has an RS232 protocol interface.
RTC: Real-Time-Clock, to keep track of the date and time of the sent and received data.

- Prediction Accuracy (outages and damage): Outputs of this model are as good as the inputs.
For accurate inputs, the model should maintain prediction accuracy 15%-20%. 20% should be
the worst case error for this model.

- Sensor sensitivity: Sensors should react to the weather conditions within the error range
specified above. Temp sensor’s 1°C error, Humidity sensor’s +/- 2% error, Anemometer’s +/-

- Converter precision: It should convert AC voltage to DC voltage with an accuracy of 6%.

Testing Procedures:
Although our project is primarily a software oriented project, we will still have to complete tests
that ensure our individual sensors, and completed weather station is working correctly.
Modular Testing:
1) Temperature Sensor
 We will test the temperature sensor by connecting it to a logic analyzer and comparing the
results with a digital thermometer. This test will ensure that the temperature sensor is working
properly and will demonstrate the sensor’s accuracy.
2) Wind Sensor
We will test the Wind sensor by connecting it to a logic analyzer and putting a fan on the sensor.
The fan will have three different speeds, and we will test the sensor on all of these. This test will
allow us to see if the wind sensor is behaving correctly.
3) Humidity Sensor
We will test the Humidity sensor by connecting it to a logic analyzer and comparing the results
with a digital weather device that measure humidity. This test will ensure that the humidity
sensor is working properly and will demonstrate the sensor’s accuracy.
4) A/D converter
Our circuit contains two A/D converters so we will have to test both of them. They will both be
connected to a 120V 60Hz input using the function generator for the test, then we will use the
Digital Multimeter to ensure that the one A/D converter outputs 5 Volts and the other A/D
converter outputs 24 volts.

Once we have completed testing on the individual components, we will have to run a series of
test on the assembled weather station. These tests are vital since our weather station will have to
be durable enough to work during storms and cold weather.
1) Rain Test- We are concerned how are box will react when it gets very wet on this test.
Therefore, we will simply pour water over the box, and then make sure all of the sensors are still
behaving properly even though they are wet.
2) Cold/Ice Test- In this test we are concerned about how are box will act in very cold conditions.
Therefore, for this test we will put ice on and around the box and let the box sit in the ice for at
least an hour. After the hour has passed we will check all the sensors to ensure they are working.
3) Heat Test- The purpose of this test is to ensure that the box will work in very hot conditions.
In order to check this we will use a couple hair dryers to heat the box. Once the box is heated, we
will again look at the output of the sensors to check for accuracy. If our weather station passes all
of these tests, we can be fairly certain that the product meets performance requirements
(approximately 90% certain). To be absolutely certain we would need access to a Thermal
Chamber and Wind Tunnel.

The two A/D converters that we will be building must have a limited ripple voltage to ensure that
our sensors work correctly. The A/D converter that outputs 5 Volts should have a ripple voltage
of less than two percent of the open circuit output voltage. The A/D converter that outputs 24
volts should have a ripple voltage of less than four percent of the open circuit output voltage. We
will test this by hooking one end of the A/D converters to the function generator which will have
a 120V 60Hz wave set on it. We will then put the oscilloscope on the output of the converter and
measure the ripple voltage. Even though we would like the ripple voltages to be two and four
percent, our circuit could probably handle five percent without error.


       Item                Part # Name          Price     Price     Manufacturer            Site
                                           (per 1)     (per
Wind Speed Sensor          Vortex           71.86     71.86       Audon        Audon
 Pressure Sensor          mpx4115a          10.34      7.82      Freescale    Digi-Key
Temperature Sensor       TC77-3.3MOA         1.00     0.69       Microchip    Digi-Key
 Humidity Sensor         HIH-4000-001       20.91     12.73      Honeywell    Digi-Key
   Watertight            R132-080-000       31.48     31.48      Hammond      Digi Key
   Enclosure                                                  Manufacturing
  Zener Diode             PLVA650A           0.50     0.16          NXP       Digi-Key
      (x2)                                   1.00     0.32    Semiconductors
     Diode               MA2C0290BF          0.58     0.13      Panasonic -   Digi-Key
      (x2)                                   1.16     0.26          SSG
     RTC                   BQ3285            2.10     2.85         Texas      Digi-Key
PIC Micronctroller       3500               560.00   560.00        AGM         AGM
      4 to 1           EL0052W               19.13    16.25       Airlink      Airlink
VoltageTransformer                                             Transformers Transformers
    4k Resistor        25J4K0E               1.82      1.00       Ohmite      Digi-Key
       (x6)                                 10.92      6.00
   10k Resistor        20J10KE               3.16      1.74         Ohmite    Digi-key
       (x6)                                 18.96     10.44
 100uF Capacitor    FK22X5R0J107M            3.56      1.05     TDK Corp      Digi-Key
       (x2)                                  7.12      2.10
 100nF Capacitor   B37987M1104K000           0.48      0.14     EPCOS Inc.    Digi-Key
       (x2)                                  0.96      0.28


Per 1 Product: $756.94

Per 1000 Manufactured Products: $723.08

Average of 120 hours per Engineer x 3 Engineers = 360 Hours Total
Average hourly wage of $50 per hour per engineer
Total Labor Costs = $50/hour x 2.5 x 360 hours = $45,000

GRAND TOTAL COSTS = $37,500 + $756.94 = $45,756.94
 Week      Task                                                       Responsibility of:
 Week 1    Finish introduction and design/block diagrams for          Tamer
  9/14     Project Proposal
  9/14     Create schedule and work on cost analysis for parts        Scott
           and labor for Proposal
  9/14     Work on testing procedures and testing analysis for        Michael
 Week 2    Extract oatage data from OAS                               Tamer
 (9/21)    Start working on reports and datasets using SAS
  9/21     Have all parts ordered                                     Scott
           Begin research of sesnsors
  9/21     Work on design and layout for AC to DC convertor           Michael
 Week 3    Start design for database                                  Tamer, Scott, Michael
 (9/28)    Have statistical discussions with professors
           Continue working on/debugging SAS code
 Week 4    Have database finished                                     Tamer
 (10/5)    Begin entering weather data
           Investigate Predictive models
  10/5     Begin assembling and testing of sensors                    Scott
           Investigate Predictive models
  10/5     Work on assembling A/D convertors                          Michael
           Investigate Predictive models
 Week 5    Start programming of microcontroller/software              Tamer
 (10/12)   Begin assemblance of the predictive model
           Furthur testing of sensors in different weather            Scott
  10/12    conditions
           Begin assemblance of the predictive model
           Perform testing on assembled A/D convertor                 Michael
  10/12    Begin assemblance of the predictive model
 Week 6    Integrate all circuit components for the weather           Tamer, Scott, Michael
 (10/19)   station to collect weather data
           Start testing on predictive model
 Week 7    Begin testing of assembled weather station                 Tamer, Scott, Michael
 (10/26)   Continue to test completed predictive model
 Week 8    Have project completed and ready for mock-up demos         Tamer, Scott, Michael
 Week 9    Start working on final presentation                        Tamer, Scott, Michael
  (11/9)   Continue testing for demos
 Week 10   Have final presentation power point completed              Tamer, Scott, Michael
 (11/16)   Begin preparing to give presentation
           Evaluate tests and data to confirm if they are viable or
 Week 11   Begin writing up of Final Paper                            Tamer, Scott, Michael
 (11/23)   Perform final tests and touch ups on completed
           Final evaluations of end product and its performance
 Week 12   Have completed project ready for final Demo                Tamer, Scott, Michael
 (11/30)   Have Final Paper finished and ready to turn in
Week 13   Final papers due          Tamer, Scott, Michael
 (12/7)   Final Lab notebooks due

To top