Modeling the Application

The following Class Diagram describes the application objects using the Unified Modelling Language (UML) notation.

Figure 11-7 Use Case Diagram for Power Demand Cartridge

Sample Queries

Modelling the application in this way, makes possible the following specific queries:

• Find the cell (geographic quadrant) with the highest demand for a specified time-period.

• Find the time-period with the highest total demand.

• Find all cells where demand is greater than some specified value.

• Find any cell at any time where the demand equals some specified value.

• Find any time-period for which 3 or more cells had/have a demand greater than some specified

• Find the time-period for which there was the greatest disparity (difference) between the cell with the minimum demand and the cell with the maximum demand.

• Find the times for which 10 or more cells had demand not less than some specified value.

• Find the times for which the average cell demand was greater than some specified value. (Note: it is assumed that the average is easily computable by TotalPowerDemand/100.)

• Find the time-periods for which the median cell demand was greater than some specified value. (Note: It is assumed that the median value is not easily computable).

• Find all time-periods for which the total demand rose 10 percent or more over the preceding time's total demand.

These queries are, of course, only a short list of the possible information that could be gleaned from the system. For instance, it is obvious that the developer of such an application would want to build queries that are based on the information derived from prior queries e.g.,

• What is the percentage change in demand for a particular cell as compared to a previous time-period?

• Which cells demonstrate rapid increase / decrease in demand measured as percentages greater / lesser than specified values.

The Power Demand cartridge as implemented is described in the class diagram below.

Figure 11-8 Use Case Diagram for Power Demand Cartridge

The utility gets ongoing reports from weather centers about current conditions and from power stations about ongoing power utilization for specific geographical areas (represented by cells on a 10x10 grid). It then compares this information to historical data in order to predict demand for power in the different geographic areas for given time periods.

Each service area for the utility is considered as a 10x10 grid of cells, where each cell's boundaries are associated with spatial coordinates (longitude/latitude). The geographical areas represented by the cells can be uniform or can have different shapes and sizes. Within the area represented by each cell, the power used by consumers in that area is recorded each hour. For example, the power demand readings for a particular hour might be represented by Table 11-2.

Table 11-2 Sample Power Demand Readings for an Hour

	1	2	3	4	5	6	7	8	9	10
1	23	21	25	23	24	25	27	32	31	30
2	33	32	31	33	34	32	23	22	21	34
3	45	44	43	33	44	43	42	41	45	46
4	44	45	45	43	42	26	19	44	33	43
5	45	44	43	42	41	44	45	46	47	44
6	43	45	98	55	54	43	44	33	34	44
7	33	45	44	43	33	44	34	55	46	34
8	87	34	33	32	31	34	35	38	33	39
9	30	40	43	42	33	43	34	32	34	46
10	43	42	34	12	43	45	48	45	43	32

The numbers in each cell reflect power demand (in some unit of measurement determined by the electric utility) for the hour for that area. For example, the demand for the first cell (1,1) was 23, the demand for the second cell (1,2) was 21, and so on. The demand for the last cell (10, 10) was 32.

The utility uses this data for many monitoring and analytical applications. Readings for individual cells are monitored for unusual surges or decreases in demand. For example, the readings of 98 for (6,3) and 87 for (8,1) might be unusually high, and the readings of 19 for (4,7) and 12 for (10,4) might be unusually low. Trends are also analyzed, such as significant increases or decreases in total, per-neighborhood, and per-station demand over time.