INFORMATION SHEET 20

Interim Guidance on the Product Mix and Output Algorithm

1 Background

For any site that might fail to meet the 1^st Milestone Target, one of the most important ways of avoiding loss of CCL discount is to apply the Product Mix and Output Algorithm (PMOA).

The PMOA is a methodology that allows you to adjust your target if energy efficiency for the site has got worse for one of the following reasons:

1. Output has fallen since the base year
2. Product mix has changed and you make products with higher average energy intensity than in the base year.

The PMOA is a potentially powerful tool to help you retain your CCL discount. It is "one-sided in your favour". This means that you can invoke the PMOA if changes in output or product mix make you less efficient, but the PMOA will not be used if the changes are the other direction and beneficial to the site!

The only negative aspect of the PMOA is that it is not necessarily simple to apply. The PMOA methodology requires a mathematical relationship to be established to link energy consumption with output level or product mix variation. This inevitably requires a detailed understanding of the way energy is used and regularly gathered good quality data on which to base the relationship.

2 Objective of this Information Sheet

DEFRA have recently published a guidance document on the PMOA:

CCA(01)08 Guidance on how to adjust targets to
allow for product mix and throughput changes

This DEFRA document is available for downloading from our website on www.cclevy.com. It is a rather lengthy (25 pages) and complex document. This Information Sheet is intended to summarise key aspects of CCA 08.

This Information Sheet is intended as an interim guidance note on the PMOA. With support from FDF and the Carbon Trust, Enviros are currently carrying out a number of PMOA case studies. When these are complete (by Easter 2002) we will publish more detailed guidance incorporating practical feedback from the case studies.

3 Basic principles - changes in product output

If a site makes a single product then it can be expected that energy efficiency (the energy used to make each tonne of product) will get worse if production level drops. This is logical because the "fixed base load" energy use is spread over a smaller tonnage of product. Figures 1 and 2 show how a drop in production results in an increasing SEC (specific energy consumption, kWh/tonne).

	Figure 1 shows that the weekly fixed base load of the site is 70kWh (this is the intercept of the line through the y axis). The slope of the line is equal to the rate at which energy is used during production. Here, the energy use in kWh is equal to two times the level of production (y=2x).
	Figure 2 plots the energy efficiency in kWh used per tonne, against the production level. Using figure 1, at a production volume of 40 tonnes, the energy used is 150 kWh. Therefore, the SEC equals 150 divided by 40 which equals 3.75 kWh/tonne.

Figure 2 clearly shows that if output falls from 100 to 60 there is only a small drop in the SEC, but if production fell further to 40 tonnes, a major adjustment in the target can be justified.

Using the PMOA, if you can derive an appropriate equation to link energy use with output level you will be able to make an adjustment to your target to account for a drop in production. A key requirement is to gather sufficient data to draw graphs of this type with a reasonable degree of accuracy. The scatter displayed in the graphs above, clearly demonstrates that the more data you have the more accurate the equations. In most cases this will require weekly energy and production data to be plotted. Note that the graphs above should be plotted separately for each energy source, i.e. a separate graph for electricity and gas.

4 Basic principles - changes in product mix

If you make two different products it is likely they will each have a different "energy intensity". For example in the bread industry it may take more energy to bake a "premium quality crusty loaf" than to bake a "standard loaf". If you can derive an appropriate equation to link energy use with product mix you will be able to make an adjustment to your target to account for a change in such production mix.

With 2 products it is usually possible to use a "multiple regression" analysis to obtain an equation linking energy use to variations in these 2 variables. Spreadsheets like Excel have tools to help with such analysis. Here is an example:

In Excel 95 or later, the multiple regression analysis can be undertaken by highlighting the table in the spreadsheet, and going to Tools, then Data Analysis (this is an 'Add-in' feature of excel). Choose the 'Regression' analysis and set the y range to be the gas use and the x range to include both product columns. The coefficients will be calculated and appear similar to below:

Gas Use = 807.9 + ( 8.87 x Product X ) + ( 3.91 x Product Y )

5 More Complex Situations

Unfortunately, in many situations the application of the PMOA will be much more complex than described above. You may have a combination of a drop in total output and a change of product mix. Also you may have a significant number of products each with a different level of energy intensity.

In theory, the PMOA can be applied even in the most complex situations. However, as a general rule you will need an increasing amount of energy data (in the form of sub-meter readings) as complexity increases.

For a simple output equation (for a site with just 1 product type) or a simple mix equation (for a site with just 2 products) it may be sufficient to base the PMOA algorithm on just the main utility meters. In more complex situations data from sub-meters will be vital (e.g. a sub-meter on a production line making one particular product).

6 Key steps in applying the PMOA

• Description of Product Groups;
• Description of energy use and metering on site;
• Outline of PMOA Development;
• Demonstration of PMOA accuracy.

7 Timetable for the PMOA

Now:

• Define basic requirements
• Start gathering data
• If required to fit extra meters (this is an urgent requirement if you are have time to collect sufficient data from the new meters).

March 2002:

• Review more detailed PMOA guidelines available by late March.

June/July 2002:

• Prepare submission describing your PMOA for DEFRA pre-approval.

November/December 2002:

• Complete your PMOA target adjustment calculations using a full set of data for October 2001 to September 2002.

8 Possible Approaches to Applying the PMOA

In this section we describe some possible scenarios in which the PMOA will be applied in the food and drink industry.

More detailed case studies based on these and other possible scenarios will be presented in the more detailed Guidelines to be published in March 2002.

1. Simple process, one product, no other influences
   This example was covered in section 3 earlier. The aim is to produce a single regression relationship for each fuel type (electricity, gas etc.) against output level. As there is a single product, there is no particular need for analysis of sub-metered data (providing the site meets the 90/10 rule and the total site energy consumption is relevant to the PMOA). If the opportunity arises, measuring base load during a shutdown period could be useful to help confirm the intercept of the regression graphs.
2. Two or more products, on same production line, product changes once or twice per week
   Collect weekly data of each production variable and each energy type. Produce a multiple regression relationship for each fuel type (see Section 4). Short-term data campaigns should be carried out to measure performance over batches of each type of product. These can be used to verify the weekly data.

Sub-metering could be useful if one of the products uses certain equipment not used by the other. These bits of equipment can be sub-metered to help prove the difference in energy use of the two products.

3. Two or more products, on same production line, product changes less than once per week
   Some sites make products on a seasonal basis. For example, in the food sector, a frozen vegetable plant may freeze peas for 3 months, then move on to sprouts and then to root vegetables.

In this situation, you should try to treat the data separately. Effectively the facility operates as a series of simple one-product processes. You should derive a single regression relationship for each product and make adjustments for the length of each operating season.

4. Two or more products, on same production line, product changes hourly
   This is a slightly more difficult situation, as it may be difficult to collect data for each product individually. If this proves to be the case, you must ensure data collection is weekly, not monthly (to provide more data points) and aim to use sub-meters where possible.
5. Two or more products, on different production lines
   This is a very common situation where several different products are being made in different parts of the factory. Production is usually simultaneous.

In this situation the use of sub-metering is important if you are trying to verify the energy intensity of each type of product. The main production lines should be separately metered. If possible each of the central utilities such as steam, compressed air, and refrigeration should be sub-metered for each production line.

6. Two or more unlinked parts of the factory with different variable
   In some situations a single facility may have two separate processes with different variables that would be "confused" if analysed together. In this case the use of sub-metering could be very important. Ideally the two processes of the facility will be separately sub-metered and you would develop a multiple regression analysis independently for each half.