The content of the invention
The invention provides a kind of appraisal procedure of Switching Power Supply load capacitance equivalent resistance, ESR can be accurately assessed
Change, lifting ESR calculate robustness.
The present invention employs following technical scheme to solve the above problems:
A kind of appraisal procedure of Switching Power Supply load capacitance equivalent resistance, it is characterised in that comprise the following steps:
S1. capacitance voltage and inductive current are measured, obtains N group measurement data;
S2. calculation matrix is built;
S3. the load capacitance ESR calculation of initial value initial optimization estimates and broad sense noise work(of preceding N ranks moment state are utilized
Rate gain;
S4. new capacitance voltage and inductive current calculated value are inputted, utilizes broad sense noise power gain and optimal estimating value
Calculate new equivalent series resistance assessed value;
S5. broad sense noise power gain is updated, waits new data to arrive, repeat step S4, renewal calculates equivalent series electricity
Hinder assessed value.
Further, the step S1 is comprised the following specific steps that:
S11. the data window length for selecting unbiased finite impulse response filtering algorithm wave filter is N;
S12. N groups capacitance voltage and inductive current measured value are obtained;
Further, the method for calculation matrix is built in the step S2 to be included, and utilizes the capacitance voltage of measurement, inductance electricity
Relation between stream and load capacitance equivalent resistance establishes calculation matrix.
Further, initial optimization estimate and broad sense noise power gain computational methods are calculated such as in the step S3
Under:
Initial optimization calculation formula is expressed as:
Broad sense noise power gain calculation formula is:
GN=(CTC)-1。
Further, the step S4 Computational Methods are built using new capacitance voltage and inductive current measured value
New data matrix Yk, then broad sense noise power gain matrix is updated:
The iterative calculation equation of equivalent series resistance assessed value is:
Further, the step S5 is concretely comprised the following steps:
S51. the root mean square of the assessed value deviation at K moment is recorded:
S52. the broad sense noise power gain adjustment factor θ at kth moment is calculatedkIt is defined as:
S53. wait new measurement data to arrive, repeat S4-S5 steps, update equivalent series resistance assessed value.
Compared with prior art, beneficial effects of the present invention:
(1) it is used as broad sense noise power gain adjustment factor by defining measurement data and estimate deviation, dynamic is adjusted
Whole broad sense noise power gain, realizes the adaptability to different working conditions.
(2) electrochemical capacitor ESR is calculated using improvement UFIR filtering algorithms in real time, can avoided in advance to systematic survey noise
Characteristic models.
(3) being switched fast under different working conditions is realized by using accommodation coefficient of gain.
Embodiment
The present invention is further described with reference to the accompanying drawings and detailed description:
With reference to figure 1, the invention provides a kind of Switching Power Supply load capacitance equivalent series resistance appraisal procedure, including it is following
Step:
S1. capacitance voltage and inductive current are measured, obtains N group measurement data;
S2. calculation matrix is built;
S3. using preceding N ranks moment state load capacitance equivalent series resistance calculation of initial value initial optimization estimate and
Broad sense noise power gain;
S4. new capacitance voltage and inductive current calculated value are inputted, utilizes broad sense noise power gain and optimal estimating value
Calculate new equivalent series resistance assessed value;
S5. broad sense noise power gain is updated, waits new data to arrive, repeat step S4, renewal calculates equivalent series electricity
Hinder assessed value.
Further, the step S1 is comprised the following specific steps that:
S11. the data window length for selecting unbiased finite impulse response filtering algorithm wave filter is N;
S12. N groups capacitance voltage and inductive current measured value are obtained.
Further, it is according to circuit structure that calculation matrix method is built in the step S2, utilizes the electric capacity electricity of measurement
Relation between pressure, inductive current and load capacitance equivalent resistance establishes calculation matrix.
Further, initial optimization estimate and broad sense noise power gain computational methods are calculated such as in the step S3
Under:
Initial optimization calculation formula is expressed as:
Wherein:C is calculation matrix, YNFor N groups capacitance voltage and inductive current measured value.
Broad sense noise power gain calculation formula is:
GN=(CTC)-1。
Further, in the step S4, circular is as follows:
Utilize new capacitance voltage and inductive current measured value structure new data matrix Yk, update broad sense noise power gain
Matrix:
Wherein:θkFor broad sense noise power gain adjustment factor, initial value 1.
The iterative calculation equation of equivalent series resistance assessed value is:
Wherein:For kth moment capacitor equivalent series resistance estimate,
Because when circuit is in different working condition, circuit power consumption is different, Switching Power Supply delivery efficiency is different, now line
Wave property can also have differences.In order to improve the flexibility of algorithm, the present embodiment is by defining measurement data and estimate deviation
Value is used as broad sense noise power gain adjustment factor, and dynamic adjusts broad sense noise power gain, realized to different working conditions
Adaptability.So further, the step S5 is concretely comprised the following steps:
S51. the root mean square of the assessed value deviation at K moment is recorded:
Wherein, β is the data dimension of assessed value.
S52. the broad sense noise power gain adjustment factor θ at kth moment is calculatedkIt is defined as:
S53. wait new measurement data to arrive, repeat S4-S5 steps, update equivalent series resistance assessed value.
If assessed value change of error is very big it can be seen from the formula in S51-S52, it is understood that there may be power condition changing feelings
Condition, broad sense noise power gain is now corrected, reduce assessed value error, if assessed value change of error very little, it is believed that assessed
As a result it is accurate, increase broad sense noise power gain, further speed up the convergence rate of filtering algorithm.
The present embodiment also further illustrates the good characteristic of the present invention using specific data, is Switching Power Supply with reference to figure 2,
The circuit includes control source Vin, switchs control pipe M1, diode D1, inductance L1, load capacitance C1, and load R1 is formed, and its is defeated
It is 100V, output voltage 12V, inductance L1=33uH to enter voltage, electric capacity C1=150uF, switching frequency 150kHz, circuit simulation choosing
Fixed Switching Power Supply electric capacity equivalent resistance is 200m Ω, by the dutycycle for adjusting M1 control signals, it is possible to achieve defeated to loading
Go out voltage Uo control.Needed in the appraisal procedure of computational load capacitor equivalent resistance collection flow through inductance size of current and
The voltage swing at electric capacity both ends.Inductive current measurement can be carried out using current sensor mode, and capacitance voltage measurement uses ADC
Mode is realized.
The performance comparision of algorithms of different when table 1 gives known noise characteristic under single working condition.Can be with by analysis
Find out, when known to noise characteristic, Kalman algorithm performances are optimal, and modified UFIR is achieved approximately with UFIR filtering algorithms
Performance, and be much better than directly to calculate using initial data.
Single operating mode under the known noise characteristic of table 1
Algorithm |
Average |
Variance |
Directly calculate |
199.93 |
9.72 |
Kalman |
199.42 |
2.83 |
UFIR is filtered |
200.39 |
3.45 |
Improve UFIR |
202 |
3.34 |
When noise characteristic is unknown, knot accurately can not be calculated to noise modeling, reference table 2, now, kalman filtering algorithms
Fruit fluctuation increase, UFIR filtering algorithms do not need noise statisticses modeling with modified UFIR algorithms, and its algorithm performance is kept
It is stable.
Single operating mode under the unknown noise characteristic of table 2
Algorithm |
Average |
Variance |
Directly calculate |
201.37 |
11.53 |
Kalman |
201.56 |
5.02 |
UFIR is filtered |
199.61 |
3.31 |
Improve UFIR |
200.52 |
3.61 |
Reference table 3 is complex working condition under unknown noise characteristic, the data statistics of each method.When circuit is in various working bar
Part, and when repeatedly being switched, ESR assessment result variances significantly improve.Because noise characteristic is unknown, different operating mode noises cause
The fluctuation of Kalman filter result of calculation is larger.UFIR filtering algorithms due to when operating mode switches iterative regression speed it is slower,
The variance fluctuation calculated in this period is larger.Improve UFIR and accelerate iterative regression speed, its result of calculation variance is minimum, property
Other algorithms can be better than.
Complex working condition under the unknown noise characteristic of table 3
Algorithm |
Average |
Variance |
Directly calculate |
200.01 |
12.37 |
Kalman |
200.44 |
7.11 |
UFIR is filtered |
199.74 |
5.95 |
Improve UFIR |
200.52 |
4.02 |
In the present embodiment, be can be seen that from above-mentioned several groups of emulation when noise statisticses are unknown and operating mode switching is frequent
Under the conditions of, modified UFIR filtering algorithms embody stronger robustness, have preferably adapted to real application scenarios.
Therefore, the present invention is in the assessment for being directed to Switching Power Supply electric capacity equivalent resistance based on modified UFIR filtering algorithms
The relatively other prior arts of method make great progress.
Non-elaborated part of the present invention belongs to the known technology of those skilled in the art.It is in addition, described in detail above
The preferred embodiment of the present invention.It should be appreciated that one of ordinary skill in the art without creative work can according to
Many modifications and variations are made in the design of the present invention.Therefore, all technician in the art are under this invention's idea existing
Have on the basis of technology by the available technical scheme of logical analysis, reasoning, or a limited experiment, all should be by right
Determined by claim in protection domain.