I-DEV0386: Measurement of the generation output of the HFO plant

This request for deviation is being submitted to address three issues related to monitoring of parameters for calculation of emission reductions that were raised during verification of emission reductions from the West Nile Electrification Project (WNEP) for the monitoring period January 1, 2005 to October 31 2009. The issues are discussed below.
1. Generation output delivered by the HFO Plant: The monitoring plan contained in registered PDD requires the generation output from the HFO plant to be directly measured (m)., the net generation output delivered by the HFO Plant has to be measured. However, actual implementation of the monitoring plan is that the generation output is being calculated (C) using measurements of two energy meters – one for gross generation and the other for auxiliary consumption. For this reason, along with this request for deviation for the present monitoring period, the PP is also submitting a request for revision of the monitoring plan to reflect the actual monitoring practices.
In addition to not directly measuring the net generation, the HFO plant did not have a meter to monitor the auxiliary consumption for the period of 1st January 2005 to 7 January 2007 (date of installation of the auxiliary meter (reference: W0500036 CT)) and only the gross electricity output was measured. In order to meet the data gap, the average daily auxiliary consumption that was monitored and recorded subsequently during the period 7 January 2007 to 31 October 2009 has been applied for the period for which the auxiliary consumption data was not monitored. This is a reasonable approach as it uses average auxiliary consumption over a period of 34 months, which is long enough to represent actual auxiliary consumption for all loading conditions.

The average auxiliary consumption over the period of 34 months (period during which the auxiliary consumption has been monitored) works out to be 0.8791 MWh/day ( 5.53% as percentage of gross generation) based on the actual data monitored.
In addition, for the purpose of conservativeness, the PP proposes to apply the percentage error of 0.09%, that has been reported to be the calibration error of the auxiliary meter (W0500036 CT) – conducted on 30 July 2008 – for determining the auxiliary consumption for the period of pre installation of the auxiliary meter. In other words, the auxiliary consumption is estimated based on the measured auxiliary consumption that has been further adjusted for the calibration error (0.09%). With this adjustement, the auxiliary consumptionis 0.8799 MWh/day.
An auxiliary consumption of 0.8799 MWh/day, when expressed as a percentage of gross generation, works out to 6.3%, which is considered to be conservative given the designed auxiliary consumption of the diesel plants is reported to be normally in the range of 1-2% .
We are hereby requesting the approval of the deviation that the net generation is calculated from the gross generation and auxiliary consumption and that the approach of using the average daily auxiliary consumption (0.8799 MWh/day), calculated using the actual data monitored over a period of 34 months (from 7 January 2007 to 31 October 2009) and further adjusted for the calibration error of the auxiliary meter, be allowed for the period 01st January 2005 to 07th January 2007.

2. Calibration of Energy Meter: A SEPAM meter (ref. SEPAM 1000 +40), henceforth referred to as “old meter”, was initially installed in the HFO plant to measure the electricity generation (gross output) from the HFO plant. However, as this meter is a part of an overall assembly, there were practical constraints in getting the calibrations done after it was installed.

Although the old meter works fine and measures the electricity accurately, on the advice of SGS , a new energy meter (Landis + Gyr model 400, type Z D4 CT S2), henceforth referred to as the “new meter”, has been procured and installed in June 2009. A certificate on calibration for the new meter is available. The issue is on establishing the accuracy of the data on electricity generation that is monitored and recorded for the period prior to installation of the new meter, as the old meter could not be calibrated.

The “Guidelines for assessing compliance with the calibration frequency requirements” states the following:

If during verification of a certain monitoring period, the DOE identifies that the calibration has been delayed and the calibration has been implemented after the monitoring period in consideration (i.e. the results of delayed calibration are available), the DOE may conclude its verification, provided the following conservative approach is adopted in the calculation of emission reductions:
a. Applying the maximum permissible error of the instrument to the measured values, if the results of the delayed calibration do not show any errors in the measuring equipment, or if the error is smaller than the maximum permissible error; or
b) Applying the error identified in the delayed calibration test, if the error is beyond the maximum permissible error of the measuring equipment.

Since the old meter could not be calibrated, the accuracy of the old meter readings have been ascertained with the installation of the new meter and comparing the readings of both meters and hence meeting the requirement of getting delayed calibration. Since the old meter has not been calibrated (due to practical constraints), the PPs are submitting this request for deviation seeking approval for using the following alternative method to establish accuracy in-lieu of calibration.
• The PE has compared the performance of the old meter and the new meter by monitoring the generation output by both the meters simultaneously for a period of 16 days during June/July 2009, Considering the data reported by the new meter to be correct (as it is calibrated one), the error of the old meter is calculated as the difference between the data reported by the old meter and the new meter for each day of the comparison period. It is observed that in most of the cases, the readings of the old meter are lower than the new meter. The mean error over a period of 16 days is caculated to be -0.06 MWh/day (-0.32% on an average). Being on lower side, the old meter readings are considered conservative and these values are used for baseline emission calculations.
• To further establish the accuracy of the data reported by the old meter, the PP has
- calculated the difference between the daily generation data reported by both meters and analysed its variation by using standard deviation and standard error for the data noted over the same 16-day period.
- carried out t-test on data reported by both the meters during simultaneous monitoring to assess the accuracy
Standard Deviation and Standard Error Analysis:
As can be seen from the data submitted as attachment to this request (Table_1), the mean Percentage Error is calculated to be -0.32%. The Standard Error of the Percentage Error is calculated to be 0.075%. This means that on one extreme, the percentage error could be -0.394% (-0.32 - 0.074) and on another extreme, the percentage error could be -0.246 % (-0.32+0.074). The error does not become positive in either of the extremes. Therefore, the data reported by the old meter is considered accurate, and conservative compared to the new meter.

T-Test Analysis
For the purpose of this analysis, the data reported by the new meter is considered as the control group, and the data reported by the old meter has been compared with it. The 2 sample t-test method has been used to establish the accuracy - one sample comprising data reported by the new meter shows true readings (sample –Y), and the other sample comprising data reported by the old meter (Sample-X).

Applying the analysis on these 2 samples X and Y, the t-test results in t = -0.009226. For a probability of 95% two-tailed, t(15; 95%)=2.131. The results of the t-test thus confirm that the 2 sets of samples are similar and there is no significanct difference between the two. This establishes the accuracy of the data reported by the old meter, when compared with the new one.

All the above analyses confirm that there is no significant difference between the data reported by the old and the new meter. The error has been reported to be just 0.32% which is lower than the maximum permissible error of the old meter, which is ± 1.0% . Therefore as per the guidance provided in “Guidelines for assessing compliance with the calibration frequency requirements”, the maximum permissible error of
-1.0% has been applied to the data reported by the old meter for conservativeness.

Now that the PP has installed a new calibrated meter, the PP hereby requests CDM EB’s approval to allow the use of data monitored by the old meter adjusted to its maximum permissible error during the period 1 May 2005 to 21 June 2009.

3. Monitoring of diesel parameters: As per the monitoring plan of the registered PDD, density and calorific values of both HFO (primary fuel) and Diesel (secondary fuel) are required to be monitored on a monthly basis. While the PP has regularly monitored the density and calorific value of HFO during the monitoring period, it has not done the same for diesel, as consumption of diesel was very negligible representing only 4.7 % of the total fuel inputs to the plant in terms of energy. While the PP commits to monitor the diesel parameters in the same way as it does for HFO during the subsequent monitoring periods, it seeks approval to apply default values for the present monitoring report. As per the “Tool to calculate project or leakage CO2 emissions from fossil fuel combustion”, the PP has looked at other sources of data, including the National Energy Balance 2005 , the Kenya Petroleum Refinery, and IPCC , which is presented as an attachment (Table_2). Accordingly the PP proposes to use the upper values provided by either one of the 3 sources for the diesel density and NCV in this monitoring report, i.e. a density of 0.865 ton/m3 and a NCV of 43.3 GJ/ton. This would lead to conservative estimates of project emissions.

Furthermore, diesel consumption in the WNEP represents a very small fraction of the total fuel compared to HFO consumption (250,000 liters of diesel compared to HFO consumption of 6 million liters) with project emissions of 860 tCO2e as compared to 18,140 tCO2e from HFO consumption.

For subsequent monitoring periods (and corresponding reporting), the PP will monitor the diesel parameters in compliance with the monitoring plan of the PDD. The PP will, as for HFO parameters, collect samples on each diesel consignment delivered to the HFO plant and send those samples to an external laboratory to measure the density and NCV. Thus 100% of the fuel consumed at the power plant will be monitored, as prescribed in the monitoring plan (ref. PDD, table 5).

We are hereby requesting the CDM EB to allow the PP to use (i) the data provided by the fuel supplier for diesel density and (ii) the IPCC default values at 95% upper confidence interval for diesel NCV for the period 1 January 2005 to 31 October 2009.

As stated above, along with this request for deviation for the present monitoring period, the PP is also submitting a request for revision of the monitoring plan to reflect the actual monitoring practices of calculating the net generation as the difference between the gross generation and the auxiliary consumption.
The other deviations described in this request for deviation are only temporarely and for future monitoring periods monitoring will comply with the requirements of the revised monitoring plan.

>> The deviations proposed above do not have any impact on the estimates of the emissions reductions (i.e no overall increase in emission reductions) due to the following reasons.
• The auxiliary consumption during the non-monitoring period (1st January 2005 to 7 January 2007 ) has been considered to be 6.3% against 5.53% observed during the period in which the auxiliary consumption was actually monitored.
• Even though the accuracy of the old meter is within the permissible limits, a maximum permissible error of -1.0% as specified by the manufacturer has been applied to the data reported by the old meter.
• Conservative values for both the density and calorific value of diesel have been used for project emission calculations.