principles.txt

		
		PROLEGOMENON TO ANY FUTURE CALCULATOR
		

Introduction
============

The 2050 Calculator is an approach to modelling the energy system
characterised by a certain philosophy. Among its design principles,
not necessarily all of which are perfectly observed in the present
version of the UK Calculator, are:

- A strictly modular structure. 

    The Calculator is made up of sectoral models. Compartmentalisation
    of the sectoral "mini-models" meant that each can be understood in
    isolation of the others.

- A focus on simple designs for the sectoral "mini-models"

- Non-cyclic information flow.

    There are no feedback loops (this is related the to point about
    simple designs). This may not strictly be considered part of the
    design philosophy, but it certainly simplified things.

- A focus on questions of the form "if this happens, then what
  happens?" rather than simply "what will happen?"

    This is perhaps the largest difference with other models, the
    Calculator does not attempt to forecast, or to optimise.

- A source trail for assumptions (à la Wikipedia's citations).

    The idea is that one admits assumptions only if they are
    facts. Eg, it is a fact that the following assumptions were
    published by such-and-such an organisation. (Or, eg, it is a fact
    that France's nuclear build rate was such-and-such per year.)

    Presently honoured more in the breach.

- A willingness to break any of the above in the interests of getting
  more people to have a better understanding of the energy system.

- An inviolate principle that it had to "add up".

    By this was meant, in part, that you can't create energy from
    nothing: the model enforced that particular physical
    principle. One way of saying this is that we simply built an
    acccounting tool for the energy system. (The rest of what was
    meant by this is the statement that, eg, total emissions are the
    sum of emissions from each sector.)

The question in this note is, can we write down what these actually
mean for an energy Calculator?


An informal sketch of a more formal model
=========================================

The following remarks are an informal sketch of what a formalisation
of a Calculator might look like.

1. We first generalise slightly the concept of an annual energy
   balance table. An annual energy balance table, as reported in
   DUKES, or the UN International Recommendations for Energy
   Statistics, is roughly, for one year:

    i.  A set of sectors;

   ii.  For each sector, a set of energy flows, one for each fuel type;
        such that the sum of flows for a sector is zero. 

   An energy balance table is used to capture historic data. What we
   would like to do is capture future possible flows. We define for
   each sector a set of possible flows: those flows (of each fuel)
   which are allowed.

   The idea here is that each sector is not a single set of flows, it
   is a "single-year-mini-model" which describes how that sectors
   inputs and outputs are related to each other.

   Thus, a "model balance" is:

    i.  A set of sectors;

    ii. For each sector, a set of input and output fuels together with
    	a subset of the space of all possible inputs and outputs. That
    	subset is the set of all allowable flows.

2. All "model balances" for a single year are thrown together. Then
   the various inputs and outputs are classified by the type of
   fuel. Finally, the system should "solve" the resulting system:
   finding that total energy balance which satisfies the conditions of
   all the model balances.

3. However, there is a trick. Fuels types are not a non-overlapping
   collection. Instead, there is a hierarchy of fuels. This is
   important, because sectors produce and consume fuels at different
   levels of the hierarchy. For example, in the present version of the
   Calculator, most sectors which burn solid fuels in fact burn "solid
   hydrocarbons". But sectors which produce solid fuels tend to
   produce "coal", "solid biomass", etc.

   This was a nice feature of the Calculator -- it simplified a lot of
   things. It meant, for example, that instead of petrol cars, diesel
   cars, and biofuel cars, ... we simply had cars with an internal
   combustion engine. (The technology is not significantly different.)

   More carefully: if a sector wishes to consume a fuel of type A, it
   should be happy to consume any fuel that is a subtype of A; a sector
   which produces a fuel of type A can be taken to have produced any
   fuel of which A is a subtype. 

   So the "solver" part of the New Calculator will need to promote and
   demote fuels as necessary.

4. Lots of questions arise: for example, is it possible to decide, for
   a given collection of model balances, whether the system as a whole
   has a unique solution?

5. 

So now the idea is this:
   
   i.   There are a collection of sectoral mini-models.

   ii.  The job of each mini-model is to create a set of model
   	balances, one for each year.

   iii. 







1. An annual model is a set of fuels, 

An energy-balance-relationship is a relationship 

1. Sectoral mini-models generate annual energy-balance-relationships. 

2. The total of all energy-balance-relationship

1. Flows (have in and out)

2. Hierarchies of flows:
   i. Fuel type
   ii. Years
   iii. Sectors

[ One of these is not like the others? Perhaps not so, actually: 

  - if produced as Lignite, may be taken as Coal; if Coal required,
  Lignite is allowed. 
  
  - if produced in 2001, may be taken as produced in 2000-2005; if
    required in 2000-2005, then 2001 is allowed. 
]
 
3. "Balances" [ need a better name for this ] 



With the above remarks in mind, we define a Calculator in terms of
three different entities: 

1. A set of "sectoral mini-models."

2. A set of "fuels". (Which include more than physical fuels; see
   later.)

3. A set of 



Mini-models
-----------