In practical experiments, it was discovered that forcefully closing Lighting channels during periods of high fee rates could inadvertently exploit a type of transaction pinning. This method involved delaying the force closure of a channel by broadcasting a low fee-rate, revoked commitment transaction, which went unnoticed by the counterparty's node due to its reliance on compact block filters. These filters, as used in LND implementations, fail to track mempool transactions and only recognize revoked commitments once they appear in blocks. This characteristic suggests that without package relay capabilities, LND's fee-bumping mechanisms could become ineffective against transactions with equal or higher fees.

Further investigations highlighted the possibility of exploiting these vulnerabilities by automating attacks during fee spikes. Such conditions are ripe for manipulation as they often forecast sustained high fees over extended periods, making it easier to estimate future fee rates accurately contrary to claims of the difficulty in predicting feerates over multiple blocks. The practical tests conducted demonstrated that under specific conditions, exploiting the Lightning Network's operational mechanics is not only feasible but can be systematically orchestrated based on predictable fee rate trends.

For more detailed insights into these experimental findings and their implications on the security of blockchain transactions within the Lightning Network, one can refer to the extensive work and research published by Peter Todd. Through his exploration, a deeper understanding of the nuances in blockchain transaction management and the potential for targeted attacks becomes evident, calling for continued vigilance and improvement in cryptographic transaction protocols.