Measurement error !?

If you measure a lot, you measure crap , as the old saying goes. Actually, it should be: If you measure wrong, you measure crap.

Especially in times of increasing demands for energy efficiency, but also in many other areas, the measurement of relevant values is indispensable. Because what is not measured can neither be evaluated nor improved. Faulty measurement results are the worst-case scenario for measurements. But what causes measurement errors? In this newsletter, we take a look at measurements in electrical energy technology and the factors that can distort the measurement result.

First: Define your expectation

Imagine that you are running a marathon and after crossing the finish line you look at your stopwatch. Is your time good or bad? The answer depends on your expectations. If you have not defined these in advance, you can neither be satisfied nor disappointed – you had no goal.

The same applies to measurements. An accuracy requirement must be defined in advance for each measuring point.

A few questions need to be answered first:

• Is the measurement relevant for billing?
• Is the measurement required for the verification of a standard?
• Are the measured values used in an energy data management system?
• Are the measured values required for a control system?
• Is there a standard for my measurement task?

If none of the above questions are relevant, the question still remains: What accuracy do I want to achieve?

The cost of a measurement is directly related to the accuracy. If the highest accuracy is defined across the board, this is reflected in the cost of the measurement. If you avoid typical errors when setting up the measurement, you can achieve a sufficiently high measurement accuracy with standard measurement technology.

Measurement via current transformer

In most cases, the current cannot be measured directly. In such cases, current transformers must be used. These convert the currents to be measured into directly measurable values. Put simply, a current transformer works physically like a small transformer. However, the selection and connection of this inconspicuous component has great potential for error. The selection of the current transformer has a direct influence on the accuracy of the measurement result.

What needs to be considered here:

1. The size of the primary ratio of the current transformer must be adapted to the expected current. The expected current should be in the last third of the primary ratio. Example: A current of approx. 220 A flows at the measuring point. A current transformer with a primary current of 300 A would still have a reserve of 80 A and the measurement would be in the last third (200 – 300 A) of the measuring range. If, on the other hand, a 1000 A current transformer were used, the measurement error would be approx. 50 % higher due to the current transformer’s excessively large measuring range. This would result in inaccurate measurement of energy quantities, particularly during off-peak periods.
2. The class of the current transformer is directly related to the accuracy of the measurement. The class indicates the maximum amplitude error at the rated current of the primary value. A class 1 transformer has a 10-fold higher measurement error than a class 0.1 transformer. Class 1 current transformers are usually used. However, if transformers are retrofitted, cable conversion transformers are usually used. This type of transformer usually has a lower class of 3 or even only 5. Current transformers also have a phase error, which also depends on the class.
3. The load of the current transformer is another point to consider. In technical jargon, this means The current transformer must neither be “overloaded” – nor “underloaded”. In simple terms, the burden is the maximum ohmic resistance that may be connected to a current transformer. The connected load can be a measuring device and the required connecting cable between the measuring device and the current transformer. If the specified burden of the current transformer is not observed, measurement deviations will inevitably occur. The burden of a current transformer is specified as apparent power in VA. This load must not be exceeded (overloaded) and should not fall too far below it (underloaded). In contrast to analog measuring devices, current digital measuring devices have a very low load.
4. The cross-section of the connecting cable from the current transformer to the measuring device depends on the cable length and the load of the current transformer. If the cross-section is too small, a significant measurement error is unavoidable.

Measurement via a summation current transformer

If the measurement task requires the use of a summation current transformer, calculations of the burden are also necessary here. Not only the loads of the current transformers must be taken into account here, but also the load of the summation current transformer and its power loss. Often the current transformers do not even have the power loss of the summation current transformer.

Accuracy of the measuring device

Each manufacturer specifies the accuracy class of its measuring devices as a quality feature. The classification indicates the maximum expected deviation of a measured value from the true value of the quantity to be measured. It is given in % and is the inherent deviation over the standardized measuring range. For analog measuring devices, this is easy to calculate.

Example:

An analog ammeter can measure a maximum of 100 A and has a class of 1. If 100 A is measured, the maximum deviation of 1 % is observed. If only 25 A is measured, the deviation is already up to 4 %.

But how does the error behave with digital measuring devices that do not have a definitive scale? Here it is more complicated and the question can no longer be answered in general terms. If a measuring device measures according to the standard, for example DIN EN 61557-12, the maximum errors are specified at a defined ambient temperature, humidity and defined harmonics. If the measuring device is operated outside the specification, the accuracy may also change. In addition, different measurement and operating ranges must be specified in the individual measured values. For example, in the current measuring range, the specified error range remains the same if the measured value is between 10 % and 120 % of the nominal value.

Example: Current transformer 500/5 A – here the specified accuracy of the measuring device is maintained within the range of 50 to 600 A. At currents below 50 A, for example during off-peak periods, a larger error may occur.

This range is specified in the voltage between 20 % and 120 %. This makes it clear why measuring device manufacturers specify several classes for different measured values.

Conclusion

As you have read, there are several sources of error that can lead to measurement errors. If you avoid the errors mentioned here when planning and setting up a measurement, you will obtain measurement results with a small and therefore acceptable deviation from the actual value. Errors are always indicated with ±. This means that both too high and too low values can be measured. In the best case, measurement errors cancel each other out.

If measurement errors occur in practice, the reason is not always incorrectly dimensioned current transformers. Incorrect wiring of the measuring device is often the reason for the incorrect measured values. In this case, the measurement error is very high and yet is often not detected.

As described in the previous text, sufficiently high measurement accuracy can be achieved with standard measurement technology. Measurements at KBR confirm this in practice. For our energy management in accordance with ISO 50001, we operate our own energy data management system, visual energy. Among other things, the electrical energy purchased from the energy supplier is measured in order to check the electricity bill.

The accuracy of our own measured values is verified by comparing them with the values from the energy supplier’s meter. As a billing-relevant measurement, the accuracy of the energy supplier’s meter can be classified as very high. In the monthly comparison of the energy supplier’s kWh measurement with the values measured by the KBR meter, the deviation is less than 0.1%. Measurements are taken using the standard multimess F96 measuring device from KBR and class 0.5 current transformers.

Our sales engineers and product management team will be happy to answer any questions you may have about your measuring task.