(Reposted)
Digital filtering technology We hereby digital filtering technique refers to the processing of eliminating interference on the acquired data in the software. In general, in addition to taking anti-interference measures in the hardware, digital filtering is performed in software to further eliminate a wide variety of interference attached to the data, enabling the collected data to be true. Reflecting the actual situation of the scene. The introduction here is a general digital filtering technology that can be used in industrial control software, which can meet the general data processing needs, and more complex digital filtering is not further introduced. 1. The signal from the dead treatment from the industrial site often continues to fluctuate within a certain range, or has a higher frequency, and the unparamids of energy are superimposed on the signal. This situation often occurs in the application Control board. Card, at this time, the last bit of the last position of the valid value of the data collected is difficult to stabilize. In this case, the deadband processing is taken, and the value of wave stop fluctuations is processed, and only changes the value when the change exceeds a value. For example, when programming, you can first divide the data at 10, then tighten, remove the fluctuation. 2. The arithmetic mean method is YK = (x1 x2 x3 ... xn) / N, sampled at different time points in one cycle, then finds the average value, this method can effectively eliminate periodic interference . Similarly, this approach can also be promoted to be averaged for several cycles. 3. The principle of the medium-value filtering method is to sort the variable values of the collected number of cycles, then take the value of the sequence worth the intermediate, which can effectively prevent sudden pulse interference Data enters. When actually use, the number of sorted cycles should be appropriate. If it is too small, it may result in removing the effect of disturbance, the number of options is too large, which will cause the sampling data to be too large, resulting in deterioration of system performance. 4. Low pass filtering method is YK = Qxk (1-Q) YK-1 cutoff frequency is f = k / 2πt. This filtering method corresponds to the captured data through a low pass filter. From the line site, it is often 4 --- 20mA signal. Its change is generally relatively slow, and the interference generally has burstic characteristics, the frequency of change is high, and the low-pass filter can be filtered out of this interference. This is the principle of low-pass filtering. When actually use, choose a reasonable Q value, and you can't achieve the goal. Its limitations, all signal mutations are seen as interference. However, this approach can be applied to some relatively special occasions, and the corresponding data processing procedure is also changed, such as the parameters of the PID. The formula of the sliding filtering method is: Yn = Q1XN Q2XN-1 Q3XN-2, where Q1 Q2 Q3 = 1 and Q1> Q2> Q3. In actual use, it may not only use one method, but in combination, the above method can be used, such as in the median filtering method, adding average filtering, to improve the performance of the filter. All in all, to flexibly select according to the situation in the scene. (www.eenginerarea.com)