10 software filtering methods
1. Limited Filtering Method (Also known as program judgment filtering method) A, method: Determine the maximum deviation value of the two sample allowed by the experience (A) determine the maximum deviation value of the two samples (a) each detected the new value: if this value is The difference in the last value <= A, this value is valid if this value is different from the last value> A, the value is invalid, give up this value, replaced this value B, the advantages: Can effectively overcome pulse interference C due to accidental factors, the disadvantage cannot suppress the periodic interference smoothness difference 2, the median value filtering method A, method: Continuous sampling N times (N-Qiqi) Press the N times sampling value The size of the intermediate value is this effective value B. Advantages: It can effectively overcome the temperature of the fluctuations caused by accidental factors, the change in the temperature of the liquid level has a good filtering effect C, Disadvantages: For traffic, speed, etc. Quick change parameters are not suitable
3, Mode Average Filtering Method A, Method: Continuous N Sampling Values When the arithmetic average operation N is large: the signal is less smooth, but the sensitivity is lower than the n value: the signal is low, but the sensitivity is more Selection of high n values: General traffic, n = 12; pressure: n = 4 b, advantage: It is suitable for filtering the signal that typically has random interference, there is a mean, and the signal is in a certain value range. Upper and down fluctuations C, Disadvantages: Real-time control of measurement speed or fast data calculation is not applicable to compare RAM 4, and the average filtering method (also known as sliding average filtering method) A, method: Take N The sample value sees that the length of a queue queue is fixed to n to sample a new data to a new data, and throw away the primary data of the original team. (Advanced First Principle) Mode average N data in the queue If the operation, the new filter results can be obtained: traffic, n = 12; pressure: n = 4; liquid surface, n = 4 to 12; temperature, n = 1 ~ 4 b, advantage: Periodic interference There is a good inhibitory effect, smoothness is suitable for high-frequency oscillation system C, Disadvantages: The inhibition of sensitivity low to accidentally occurring pulse interference is less likely to eliminate the sampling value deviation caused by pulse interference is not suitable for pulse interference More severe occasions compare RAM 5, medium-bit value average filtering method (also known as anti-pulse interference average filtering method) A, method: equivalent to "median value filtering method" "arithmetic average filtering method" continuous sampling N data Removing a maximum value and a minimum value and calculates the selection of the arithmetic mean n value of N-2 data: 3 to 14 b, the advantage: the advantages of two filtering methods can be eliminated for occasional pulse sexual interference Due to the sampling value of the pulse interference C, disadvantages: the measurement speed is slow, and the arithmetic average filtering method is more wasteful RAM
6, limited average filtering method A, method: equivalent to "limited filtering method" "Recursive Average Filtering Method" each sampled new data is limited to limit, and then send the queue to perform the average filtering process B , Advantage: The advantage of combining two filtering methods For occasional pulse sexual interference, the sampling value deviation c, disadvantage, due to pulse interference: more waste RAM7, first-order hysteresis filtering method A, method: take a = 0 ~ 1 This filter result = (1-a) * This sample value a * last filtering result B, Advantages: Periodic interference has good inhibitory effect suitable for the high fluctuation frequency C, Disadvantages: The degree of phase lag, the degree of sensitivity low lag depends on the A value size of 1/2 of the filter frequency above 1/2 of the sampling frequency 8, weighted average filtering method A, method: is an improvement in the average filtering method of the recursive, ie Data from different time is usually, the closer the data, the greater the right time. The larger the weight of the new sampling value, the higher the sensitivity, but the lower signal smoothness is B, the advantage: the system C, shortcomings suitable for shorter and sampling cycles with large pure lag time constant: for pure lag The time constant is small, the sampling cycle is longer, and the slowed signal cannot quickly respond to the severity of the system, the filtering effect is poor.
9, Wash Filtering A, Method: Set a filter counter compare each sampling value with the current valid value: If the sample value = current valid value, the counter cleans if the sample value <> Current valid value, counter 1 And determine if the counter is> = upper limit n (overflow) If the counter overflows, the current valid value is replaced, the countercer b, the advantage: There is a good filtering effect for the slowed measured parameters, which can be avoided Definitive on / off in the critical value of the controller, the value shake C, Disadvantages: For rapid change parameters, if the sampled value of the sample overflows in the counter exactly the interference value, the interference value is active as effective Value import system
10. Limited Wave Filtering Method A, Method: It is equivalent to "limiting filtering method" "blast filter method" first limit, then shake B, advantage: inherit "limited" and "shake" advantages Improved some of the defects in the "Wash Filtering Method" to avoid importing the interference value into the system C, disadvantages: the parameters for rapid changes are not suitable
