EBSILON®Professional Online Documentation
Reliable measurement values are essential key factors for a proper and economical operation of power plants. During the process analysis, systematic and stochastic deviations of these measurement values makes it difficult to prepare concluding mass and energy balances for the individual components and the entire plant. Statistical methods help to localize changes in the measurement values and thus in the equipment, in order to initiate specific maintenance measures. Data Validation means the determination of a consistent data set for a stationary operation point (pressure, specific enthalpy, and throughput) by methodical evaluation of predetermined measurement values. In EBSILON®Professional, as compared to the guideline VDI 2048 Sheet 1 (Measurement uncertainties in acceptance measurements at technical power plants, 2000, Beuth-Publishing), there is the restriction, that only the specification values provided in the system itself (measurement values and component data) can be validated. Measurement values, for which no measurement value type is defined in EBSILON®Professional, but which are included in the equations system to be satisfied (auxiliary conditions as per Section 5.2 of the VDI-Guideline Sheet 1), are set on any randomly available measurement value type. The correct use is ensured by an indexed access to the constraint.
On the basis of minimizing the deviations of the measured or the specified values from their estimates, the data validation returns the most probable process state for the real values in an over determined system. This state is described by a set of validated values, whose difference from the input values represents the probable deviations.
The validation of measurement data of a power plant is difficult because the plant model contains non-linear equations. Thus EBSILON®Professional solves this iteratively with a special Newton-iteration. From the error calculation point of view this is permitted if the partial derivatives of the values can be considered to be constant in their application range (see Section 4.4.1 of the VDI-Guideline Sheet 1, where the prerequisite for the linearization is shown). This numeric method in EBSILON®Professional is also used in the simulation mode (solution of the specific equations system).
Moreover, in both calculation types the same component physics (including chemical processes, which affect the fluid composition) is used, so that the solutions match in the redundancy-free case (no superfluous measurement values) i.e. the validation solution is then included in the simulation solution.
By default, the mass flows, pressures, and specific enthalpies on all lines are used as variables in EBSILON®Professional, which is adequate for calculating the Water-/Steam-circulation. EBSILON®Professional, however, enables the mapping of flue gas path and the combustion processes. In this case the calorific value and the material fractions (chemical composition) on all lines can be included in the calculation as additional variables. For instance, if in the scope of an acceptance measurement fuel, flue gas, ash and slag analysis are available with their confidence intervals, they can be fully involved in the calculation.
Starting from validated data, this system can be used to run consistent comparative calculations (what-if calculations) in the simulation mode for optimization, because of the identity of constraints and simulation equations. Moreover, special characteristic parameters (e.g. power loss characteristic, CHP-power fraction) can be determined only by comparing the currently validated process state with a comparative simulation solution. However, a necessary prerequisite for this is the system consistency.
To validate data in the off-design mode, an adequate modeling of the component physics is necessary. EBSILON®Professional makes use of multi-dimensional characteristic areas for mapping the off-design behaviour of heat exchangers, condensers and turbines (identification with EposNeuro). The uncertainty of the characteristic area is included in the confidence interval for the specification value, which can be validated. A high accuracy of the model is necessary, otherwise the difference between the measured and validated value can contain a considerable model error. The limitation of the component physics to mass and energy balances dispense with the plant specific physics (strict laws or approximate formulas with estimated value that can be validated) and thus reduces the equation redundancy.
In EBSILON®Professional a validated overall solution (pressure, throughput, specific enthalpy for all lines) is calculated, while maintaining all component-specific equations. The non-conformities resulting from the redundancy disagreements of the measurements, result as the differences between the measured values and the validated values at minimum error square sum. Considering the error propagation law in accordance with VDI 2048 Sheet 1 Section 4.4, estimated values, true to expectations for the deviations of the validated values, can also be calculated.
The next chapter describes the Possibilities for validation in EBSILON®Professional.