Why are the maximum values in the graph different from the maximums in the report?

The short answer is that the filtering of the graph data is different from the filtering used in the report.

What is filtering?

The data acquired from the engine shows a series of oscillations, noise, and vibrations that are inherent to combustion engines. The vibrations and noises we feel in the testing room are also felt by the data acquisition system.

This means that when measuring, we get readings with many peaks and valleys. The magnitude of this noise will depend on several factors such as mechanical imbalances, mechanical damping of the engine, engine performance, levels of electrical interference, and also the configuration of the acquisition system in terms of filtering that noise in the electronics and the way the signal is processed.

The Accudyno system produces minimal filtering of the incoming signal. We do this to have as much information as possible and also because the Accudyno system allows for obtaining precise data.

NOTE

An extremely smoothed and filtered signal is a symptom of a poor measurement system. It's important to have very rich raw data.

Analogy

If we are in a meeting with many people talking but we are interested in what the person in front of us is saying, our brain will filter out the rest of the conversations to focus on our speaker. But if in the middle of that conversation we hear someone say our name, we will probably turn to see who is calling us.

When we have a system that can capture a lot of information in great detail, we can find information that interests us in that data, whether it's the main signal -Power- or some strange signal that is being added such as vibrations, noise, or small failures.

In some systems, that background noise is completely eliminated, and we are left with a clean "conversation", but we are missing what is happening around us, and if someone calls us, we won't hear them.

In our system, that background noise is recorded, as it would be recorded with a microphone present at the meeting. However, we have tools to isolate the signal of interest from the rest of the noise. This is what filtering consists of, which is actually post-filtering since it is filtering that is done after the signals have been acquired. Just as we would change our focus in a recording of the sound of that meeting, sometimes we can pay attention to the conversation and when listening to it again, we can pay attention to the background.

PRACTICAL

In practical terms, this means that we normally want to see the signal with high filtering, since we are interested in the general power trend, not the small variations. But it's also good to see the graph with low filtering to detect small failures, vibrations, imbalances, etc. that in some cases reflect underlying problems in the engine or in the room.

Effect on the maximum

The system has two filtering levels. One is a fixed level filtering that is pre-selected for the Report and the other filtering level is selectable by the user for the Graphs.

EQUIVALENCE

In version 3.5, filtering levels range from 0 to 5. The fixed filtering of the Report corresponds to level 2.

In version 3.4, the possible filtering levels are low, medium, and high. The Report would have medium filtering.

If in the Graph we use the same level of filtering as in the Report, then the maximum peak values will be equal. But if the filtering of the Graph is higher than the filtering of the Report, for example, if we have the Graph on High - and the Report is fixed on medium- we will find that the peaks of the Graph will be smoothed more and probably the maximum will be lower than the maximum that appears in the Report.

This is more noticeable in signals with a high level of noise or vibrations. In these cases, the difference in the maximums of the filtered signal and the unfiltered signal will be more noticeable. For example, in this image we see the same signal with two different filtering levels applied. The blue signal is the raw signal, with much more information, but also with much more noise. The red signal is filtered. We see that the noise peaks of the blue signal make the maximum reach a higher value than the maximum of the red signal.

Solution

It's not strictly a solution, because there isn't a problem. They are simply different ways of viewing the same data.

But at a practical level, if we want to avoid giving explanations to clients, what needs to be done is to find the Graph filtering level that gives us the same maximum power as the Report delivers

The most practical approach would be to configure the Graph presets to have a setting with high filtering, and another with medium filtering to switch them easily with the preset buttons from 1 to 9.