You need to  establish  a new standard curve with LOQ is less than 30 ng/g .You can  either use signal to noise ratios or a standard deviation method  to establish a new LOQ.

At same time how do you know your sample has 30 ng/g  though a series of dilution ?

I think, standard addition method is applied when the concentration of the analyte in sample is relative low (below LOQ), so it needed to add standard in sample so the concentration is >LOQ. Did you develop a linear regression curve of the added standard (X axis) Vs Absorbance/peak area (as Y axis)? Was the intercept of the regression curve with X axis was - 30 ng/g? If it is so than it is correct (concentration in sample was 30 ng/g.

See atachment

Tank you Mirajkar

Tank you Indrayanto, yes we used intercept of the curve. So what i have to do for LOQ? Is it 40 ng/g or we need to add another spiked sample below than 30 ng/g and develop new regression curve?

I think if you made calibration curve that consisted of respond detector (Y axis) and added standard (X-axis) with 4-5 levels, and each level should be > 40 ng/g, if it is so, than your calibration using standard addition method is OK. Your method have LOQ 40 ng/g, it means, you cannot do quantification below 40 ng/g, that is why you used standard addition method; by this standard addition method you found that the concentration of your analyte in sample is 30 ng/g. How many replication you did  by this method (standard addition method)? If you have done  with ca 5-6 replications than RSD